(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Transactions of the American Society of Civil Engineers"

;: ;,'J}.>,(!' ;a'v:-,;, • 









TRANSACTIONS 



AMERICAN SOCIETY 



CIYIL ENGINEEKS 



(INSTITUTED 1852) 



VOL. LXIII 

JUNE, 1909 



Edited by the Secretary, under the direction of the Committee on Publications. 

Reprints from this publication, which is copyrighted, may be made on condition that 

the full title of Paper, name of Author, and page reference are given. 



NEW YORK 

PUBLISHED BY THE SOCIETY 
1909 



"^Qlt.Q, 



Entered according to Act of Congress, in the year 1909, by the American Society of 
Civil Engineers, in the Office of the Librarian of Congress, at Washington. 



Note. — This Society is not responsible, as a body, for the facts and opinions advanced 

any of its publications. 



CONTENTS 



PAPERS 

No. PAGE 

1099 FOUNDATIONS FOR THE NEW SINQER BUILDING, NEW YORK CITY. 

By T. Kennard Thomson 1 

Discussion : 

By O. F. Semsche 21 

Eugene W. Stern 24 

Edwin S. Jarrett 26 

T. Kennard Thomson 28 

HO) THE LOW STAGE OF LAKES HURON AND MICHIGAN. 

By C. E. Grunsky 31 

Discussion : 

By H. M. Chittenden 48 

C. E. Grunsky 51 



is 



1101 THE FLOODS OF THE MISSISSIPPI DELTA: THEIR CAUSES, AND SUG> 
GESTIONS AS TO THEIR CONTROL. 
By William D. Pickett 53 



1102 ELECTRIC RAILWAYS IN THE OHIO VALLEY BETWEEN STEUBEN= 
» VILLE. OHIO, AND VANPORT, PENNSYLVANIA. 

^ By George B. Francis '. 73 

W Discussion : 

C By F. Lavis 91 

V- George B. Preston 96 

J. Martin Schreiber 96 

William J. Boucher 98 

George B. Francis 99 



0^ 



1103 NICKEL STEEL FOR BRIDGES. 

By J. A. L. Waddell 101 

Discussion : 

By Charles Evan Fowler 300 

M. F. Brown 301 

H. P. Bell 302 

> L. J. Le Conte 305 

^ W. K. Hatt 306 

Of John C. OsTRcp 308 

RT. Claxton Fidler 312 

Robert E. Johnston 315 

^.yL..,^ Albert Lucius 315 

. G. Lindenthal 316 

J^ Henry S. Prichard 317 

^^^i-A*^^ Henry Le Chatelier 322 

A. Ross 324 

>^ L.Dumas 327 



isQas" 



IV 

No. PAGE 

1103 Discussion continued. 

Victor Prittie Pkrry 328 

W. H. Warren 331 

William R. Webster 337 

William H. Brkithaupt 339 

E. A. Stone 341 

C. CoDRON 342 

W. W. K. Sparrow- 345 

B. J. Lambert ^ 347 

William Marriott 348 

Henry Rohwer 348 

Samuel Tobias Wagner 350 

A. W. Carpenter 352 

Leon S. Moisseiff 358 

James C. Hallsted 361 

F. Arnodin 366 

Wilson Worsdell 378 

William F. Pettigrew 379 

J. A. L. Waddell 379 

1104 THE IMPROVEMENT OF THE OHIO RIVER. 

By William L. Sibert 388 

Discussion : 

By Theron M. Ripley 426 

William L. Sibert 426 



MEMOIRS OF DECEASED MEMBERS 



William Beverly Chase, M. Am. Soc. C. E . 429 

Martin William Mansfield, M. Am. Soc. C. E 431 

Mark William Schofield, M. Am. Soc. C. E 432 



PLATES 



PLATE 

I. 

II. 

III. 

I\'. 

V. 

VI. 

VII. 

VIII. 

IX. 

X. 

XI. 

XII. 



XIII, 
XIV. 

XV. 
XVI. 
XVII. 

XV III. 
XIX. 
XX. 
XXI. 

XXII. 

XXIII. 

XXIV. 

XXV. 

XXVI. 

XXVII. 



PAPER 

The Sitif^ei- Building, New York City lO'.Kt 

Views showing Progress in Sinking Caissorjs foi- the Singer Building.. 1099 

Views showing Progress in Sinking Caissons tor the Singer Building.. 1099 

Views showing Progress in Sinking Caissons tor the Singer Building.. 1099 

Views showing Progress in Sinking Caissons for the Singer Building.. 1099 

Views showing Progress in Sinking Caissons for the Singer Building.. 1099 

Views showing Progress in Sinking Caissons for the Singer Building.. 1099 

Views showing Progress in Sniking Caissons for the Singer Building.. 1099 

Record of Sinking Caissons for Ihe Singer Building, ]900-190T 1099 

Record of Sinking Caissons for the Singer Building, 1900-1907 1099 

The Singer Building, During Erection 1099 

Diagram showing Mean Annual Elevation of Water Surface of Lakes 

Huron and Michigan, and Discharge of St. Clair and St. Mary's 

Rivers, etc 1100 

Retaining Wall, Cooks Ferry; and Yellow Creek Bridge 1108 

Track View at Industry, Pa.; and Water-Cooling Tower, Steubenville 

Power Plant 1103 

Bending Tests, Nickel and Carbon Steels 1103 

Drifting Tests, Nickel and Carbon Steels 1103 

Close Punching Tests of Nickel and Carbon Steels, and Eye-Bar 

Material Tests 1103 

Tests of Long Columns 1103 

Tests of Short Columns 1103 

The Steamer Sprcu/ue and Tow of Barges 1104 

Map of Territory Commercially Benefited by Reliable Navigation, 

Ohio and Jlississippi Systems 1104 

Views of Lock No. 6, Ohio River 1104 

Views of Lock No. 2, Ohio River 1104 

Cross-Section of Foundation of Davis Island Dam, and Cross-Section 

of Dam No. 6 1104 

Cross-Section of Foundation of Dam No. 4, and of Dam No. 5 1104 

Views of Bear Trap Dam, Lock No. 5, Ohio River 1104 

Views of Several Locks and Dams on the Ohio River 1104 



PAOE 

3 
5 

9 
11 
13 
15 
17 
19 
19 
23 



81 
231 
235 

237 
247 
249 
393 

395 
407 
413 

415 
419 
421 
423 



VI 



ERRATA. 

Transact ions J, Vol. LX. 

Page 55 : Coulomb's Formula, on this page, should be changed to 

read as follows : 

^ „ / angle of re]iose\ 
P = Wh X tan .2 (45° — — = — ) . 

Transactions, Vol. LXII. 

Page 493, lines 21 and 22 : For "Owen Creek" read "Queen Creek." 
Page 508, line 3 from bottom: Take out the words "forests are" and 

insert in their place the words "deforestation is." 
Page 509, line 9 from bottom: Insert the word "million" after the 
word "thousand," so that the sentence will read: "A little while 
ago the country was told that a thousand million tons of soil are 
yearly washing from our agriculttiral lands into the sea." 
Page 512, line 8 from bottom: Take out the word "not." 
Page 540, line 19 : Take out the words "considered in the abstract." 



AMERICAN SOCIETY OF CIVIL ENGINEERS 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1099 

FOUNDATIONS 

FOR THE NEW SINGER BUILDING, 

NEW YORK CITY.* 

By T. Kennard Thomson^ M. Am. Soc. C. E. 



With Discussion by ]\[essrs. 0. F. Semsch, Eugene W. Stern, Edwin 
S. Jarrett, and T. Kennard Thomson. 



In August, 1906, the Singer ]\Ianufacturing Company awarded a 
contract to The Foundation Company of New York City, for sinking 
the pneumatic foundations of the addition to the old building on the 
northwest corner of Broadway and Liberty Street, now known as the 
Singer Tower. 

This contract was for a lump sum, on the basis of the foundations 
being carried down to 70 ft. below the curb, or approximately to the 
average depth of the top of the hardpan as shown by the borings ; for so 
much per cubic yard for everything between the depths of 70 and 75 ft. ; 
and for an additional price per cubic yard for everything below 75 ft. 
The price to be deducted in case the depth of 70 ft. was not reached 
was about one-half of the price to be added. The progress reports 
of each caisson, given in Table 2, Plates IX and X, will show the 
justice of this proportion. The contract stipulated that the work 
should be completed in 110 days on the same basis of depths. 



* Presented at the meeting of November ]8th, 1908. 



2 PNEUMATIC FOUNDATIONS 

This is not only the fairest method of letting caisson work, but it 
is the only method which should be adopted; for if the contract is 
made for a lump sum to bed-rock, either the contractor must bid 
excessively high to allow for uncertainties, or run a good chance 
of losing his anticipated profits and much more besides. This is 
especially true where the borings are what are known as wash borings, 
which scarcely ever reach rock, but stop at the hardpan. This state- 
ment applies at least to wash borings made for sky-scrapers, although 
it does not apply to some which have been made for some of the 
railroads, where the borings have reached bed-rock by the use of an 
occasional small charge of dynamite. 

Diamond-drill borings, of course, will find rock accurately, but 
only for the exact spot of the drill hole; for bed-rock here, known as 
New York gneiss, is very irregular, in fact it was found to be 6 ft. 
higher in one caisson than in another, although the two caissons were 
only 12 in. apart, each landing on a comparatively flat rock. In 
another case we landed one end of a caisson, 14 ft. long, on bed-rock 
and then had to dig 14 ft. deep at the other end to find rock. This is 
mentioned because the surface of the rock nnder the Singer Building 
is exceptionally level, as shown by the cross-sections of the caissons. 
Fig. 1. 

The size of the lot is approximately 75 ft. on Broadway and 115 ft. 
deep on the south side. 

The tower is square in cross-section and has 36 columns, all 12 ft. 
center to center, making it 60 ft. square — measuring from center to 
center of columns. These tower columns rest on 20 pneumatic caissons, 
as shown on Figs. 2 and 3, and are carried to bed-rock; Caisson No. 
48-49, the second to be sunk, was carried to rock as an exploration 
caisson. The remaining caissons are not under the tower, and were 
stopped in good hardpan. 

At first it was intended to stop all the caissons as soon as good 
hardpan was reached, but after several had been sunk, including Cais- 
son No. 30-43, which is one of the tower caissons, it was decided that 
all those under the tower should be carried to bed-rock, and the ques- 
tion arose as to what could be done about Caisson No. 30-43, which 
had already been carried 7 ft. into the hardpan and filled with con- 
crete. The contractors volunteered to tunnel under this caisson at 



PLATE I. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 




;i!!Sgs 




Tower of the Singer Building IHring Construi tion. 



PX !• UMATIC FOUNDATIONS 



- I-- 






■0,.•^v:■::.^i',_^.■?,V:^?, / 






• 1 " ? • f -^l' 22* ' ^ 







I 









1 ? 



r^fc: 



..,,..#, .. . 






:-< ^ r - 1 - 



;oiL:s:o;;:qv<;;:i||||„ 



"v'^^~^-i-V 



^mm^i' 



I ft 






►8 P '< -^ J ,~^ ^ ~ ' 






/A.:-^:fi.'^ ■-■a 



mjif.^i'v"''! 






/<?,, 



^f:l?:;:?::fe 



l^iiii-v:^c>^: 



. " I V!,liii>i: 'n'-i-"--^ ' 



^•| 'xe.'.-'^- ■^. ■■■-:. 
=■ , . -'-^ ^^ r ' - 'r , 

„.. . •■■ r-'' ■ - -- — 



4 • I'NEUMATIC FOUNDATIONS 

their own expense, so that the entire tower would rest on bed-rock, 
and this was done successfully as follows: 

When Caisson No. 29-35 had been sunk some 5 ft. below the 
bottom of No. 30-43 (Fig. 4), a drift was cut under and to the north 
end of the latter. This drift was about 4 ft. wide and 5 ft. high, and 
when it had been completed all the hardpan under the north end of 
the caisson was removed (Fig. 5), the distance from the bottom of 
the old concrete to the top of the rock being 15 ft. at this point. 

The hole under the north end was filled at once with concrete which 
was rammed up against the old base (Fig. 6). This, by the way, gave 
the opportunity to remove, from the original base, a piece of the con- 
crete, put in under air, and this, having been made quite wet in the 
first place, was found to be very hard and compact. It is well known 
to caisson men that concrete when made properly with plenty of water 
sets very quickly in compressed air and becomes very hard. 

As soon as the north end of Caisson No. 30-43 had been carried 
to rock, as before described, the adjoining caisson. No. 29-35, was 
excavated to rock, and the remaining hardpan under the south end of 
No. 30-43 was removed and the space concreted, thus completing, with- 
out any accident, probably what has been the only attempt ever made 
to undermine a pneumatic caisson (Fig. 7). Of course, if the original 
excavation had not been carried some 7 ft. below the cutting edge 
into good hardpan to which the concrete had firmly united, much more 
difficulty would have been encountered, if indeed the attempt had been 
made. 

It might be stated that the cutting edge of a caisson is rarely 
carried much below the top of the hardpan, and it should not be, for, if 
the excavation is carried below the cutting edge into the hardpan and 
the space then filled with concrete, the concrete forms an excellent 
bond with the hardpan, greatly reducing the load on the base, which 
is usually from 12 to 15 tons per sq. ft. Whereas, if the caisson is 
carried through the hardpan there would be almost no friction below 
the top of the hardpan. 

In this case the foundations were designed for 15 tons per sq. ft. 
In designing th(> foundations, it was decided that, if the wind pressure 
did not exceed 50% of the dead and live load, it would not be regarded, 
so the wind strains were not considered. 



PLATE II. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 





PiJEUMATIC FOUNDATIONS 




•tI on 




rNEtJMATiC FOTTXDATIONS 



0im w 



mm 



^m!^iiwmm 



i -ii 



■'i;iiW|;MiriilKflWMin//iiiMitoi 
ill (ii,>!ia''''liiii',|:,UiniJl'Uii[iii 



\mm 



llliliS^ 



ll 



-•iil|l;l|l)|i|'l|'l|i|l!f)liin|!lljli|i'l(;i|l| 



iiiiiiEllillliiMillilllli 



^il||i;ai!iii;ii!i(lfliii»iiMlWll) 



S f " 



a 



^siffmm^s!!S!^'i]^\iim]t\^\\\vv\r 




i|',')iini|ip 'iMMiii^' ii ')' iii;iMiiiiiii 
[ jljIlllllillillullIdH^ 



_LJ^ 



,,i.iiiriiiTii,iiiiiii 'I I 



m 



i!:ii!:!H!ii!.i.i!i,ii' 






Sill 'i&ii«iii , 

[i:llllMilllrllHIII,in»llllll,lllilllllli|l>lhl.i.i>llll|l I ' I U I 1 1 M I M 



r 22- 






^i::,i-5*.:!K'.>-»: 



illllllil 



-50- 



72- 



PLATE III. 

TRANS. AM. SOC. CIV. ENGR8. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 





PNEUMATIC FOUNDATIONS 



!ii)l!iliii!!,!;s.!)i|iil!i|!!iii!i 
111 




, „ ,,., ..m-mmm^ ^ ' m ,, rTi^iTli '11(^(1 Hl'ltji I ilJTliliVll 1 1 1 1 1 1 1 1 






OCO 



O Miii*'i«"'i"i""'iWi|"lli"WiJ!ll!ilMlmll 1 1 1 1 I I ' I M I ' 1 1 1 1 W 1 1 i I I I I I ' I ' I I I I ' ' ' M 1 1 P 1 1 1 1 




imiiii 
iliiii; 

llllli' V 

Hill 

nil 

Ml 




'"" liilllliiiBiijiS 

t piiiiiiaillB li! ill iiii. I 



o-.:.'z>;..b'-ii'o'-:o-".'-<;v'oi;0! 

n--.6-:-o-.i^-;o:;:S.-:-<>---;flr 

i-; ■0'.-:S^-:ci:: :;P-;;o -.=qVv 



^ .'-'- "6 '■ ■ ■■'<fi'.'j9-V'9'-i 



llli^'iiliiililliili'liliiiiiiiliiiiiililiiiil^^ 



Jji^. I C' ,,);:„,:. ;iliiliiiliiiilliiiiiilliiiiii(iilli 



O PNEUMATIC FOUNDATIONS 

The following are some of the principal facts relating to the Singer 
Building: 

Height of tower from bottom of caisson to 

top of flag-pole 74*) ft. 

Height from basement floor to top of flag- 
pole GOO " 

Height from sidewalk to top of flag-pole.. . 052 '' in. 

Height of tower from sidewalk to top of 

lantern 012 " 3 " 

Height of main building from sidewalk to 

roof 191 " S " 

Greatest depth of caisson below curb 92 " 

Area of each main floor 20 103 sq. ft. 

Area of each floor of tower 3 737 " ■■ 

Total area of floor 411 333 " '' or 9.44 acres. 

Plant. — The hoisting plant consisted of a four-boom derrick and 
two stiff-leg derricks, with five Lidgerwood double-drum engines, 
3 ft. 7 in. by 10 in. and two, 8| by 10 in., and one Lambert, 7i by 10 in.; 
in addition, there were four Rawson and Morrison boom-swinging 
gears. 

A platform was built on the level of the sidewalk about 15 ft. 
above the excavation. It is customary to excavate the lot to about the 
water level before commencing the caisson work proper, and the 
derrick was built so that carts could run under it on the street level. 
It was about 30 ft. square, with four masts 30 ft. high and 50-ft. 
booms. 

The compressor plant consisted of one Rand straight-line com- 
pressor with 14-in. steam cylinder, 18-in. air cylinder and 22-in. stroke. 
capable of pumping 1 294 cu. ft. of free air per minute, theoretical 
rated capacity, and one McKiernan 22-in. steam cylinder by 26-in. air 
cylinder by 24-in. stroke compressor, capable of pumping 1 474 cu. ft. 
of free air per minute with a speed of 100 rev. per min. Twin air 
receivers were used, each 41 in. in diameter, coupled, and 15 ft. 9 in. 
long. There was also a 14-in. air cooler, 14 ft. long. 

The concrete was mixed in one Ransome mixer with a capacity of 
24 cu. ft., and one Chicago mixer with a capacity of 18 cu. ft., both 
placed \mder the street platform. 



PLATE IV. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 






PNEUMATIC FOUNDATIONS 




Fig. 8. 



10 



PNEUMATIC FOUNDATIONS 



Fig. 1, Plate II, shows the ground on August 28th, 1906, with 
the contractors getting ready to build tlie platform. It will be noticed 
that they had already stored some 800 tons of pig iron on the site, 
ready for use, in addition to 400 tons of east-iron blocks, each block 
weighing If tons, or as much as TO or 80 pieces of pig iron. 




Fw. 11 



By September iTtli, the i)latf()rni had liccii partially luiiU on the 
Broadway level and the erection of tiio four-mastod derrick com- 
Dionced. Fig. 2, Plate II, a view taken from Prondway, shows a 
temporary stiif-log derrick with a gin-jxilc dci'rick in tlic rear used for 
the erection of the ruiii'-lHitmi derrick. 



PLATE V. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 





p 


■ 


1 


p 


v^ 








w^vi'''^^ 


ii 


i 


» 


IflftJbJi) 




ItH 


! '■*', ' 




I'NEUMATIC FOUNDATIONS 



n 





12 PNEUMATIC FOUNDATIO:^fS 

Fig. 3, Plate II, taken on September 29th, shows the four-boom 
derrick completed and in use, the two temporary derricks having been 
removed. 

Fig. 4, Plate II, taken on October 8th, 1906, shows the caisson for 
Column No. 50, the first to be sunk into position, being concreted 
around the shaft which is covered up to prevent the concrete from 
falling into the working chamber. It also shows the caisson under 
Columns 40-41 to the right, in place, which was the third to be sunk, 
the second caisson being behind the derrick. It also shows another 
caisson resting on the I-beams on the platform. 

Fig. 1, Plate III, taken on October 13th, 1906, gives a good idea 
of the temporary forms of 2-in. plank, tongued and grooved, and 
planed, used for the concrete on top of the caissons. As will be seen, 
the steel angles, usually 3 by 3 by § in., are not cut to length for 
each size of caisson, but are allowed to project beyond the corners. It 
also shows the air-lock on the first caisson sunk — 'that under Column 
No. 50 — the photograph having been taken two days before air was 
put on. 

The four-masted traveler to the right is one of the traveling der- 
ricks on the lot of the City Investing Building, which was under con- 
struction at the same time as the Singer Building, the writer being 
retained on both. Some interesting problems were encountered, one 
of which was to locate the cutting edges of the caissons on each side 
of the lot line. Owing to the caissons not being strictly plumb, those 
of each building encroached on or under the adjoining property. 
Fortunately for the owners, both sides were trespassers, otherwise a 
lawsuit might have given much work to the lawyers and engineers. 

Most specifications state very positively that a caisson shall not 
be out of plumb more than a certain amount, varying generally from 
1 to 6 in. ; but suppose a caisson is down 90 ft. in the ground and is 
found to be, say, Y in. or more out of plumb — in fact, the writer has 
heard of, but has not seen, a caisson which was 5 ft. out of plumb- - 
what can be done about it? Of course, a divergence of an amount 
like 5 ft. is absolutely inexcusable, but the best care will not prevent 
an error of a few inches. The only remedy is prevention — for an 
experienced man can start his caissons right and keep them so, and 
it is to the owner's advantage to attend to this from the start. If .i 
caisson gets much out of plumb and down to a certain depth, it is 



PLATE VI. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 





Fig. 1.— October 26th, 1906, Showing 
Lack of Space. 



Fig. 3.— Caisson Just Lifted off Truck. 




1 Tl 


LMa 


-^^^--m'^l'- 


'^€im 


^ — ^H Mr i?W 


ll 


\^^W- 


w 




^ 


11 







Fig. 3.— Needle Beams and Screw- 

Jacks used in Underpinning 

Old Buildings. 



Fig. 4. 



-April 5th. 1907, Steelwork 
Being Erected. 



PNEUMATIC FOUNDATIONS 13 

not only impossible to plumb it again, but it is also impossible to 
prevent it from getting rapidly worse. The best superintendents rely 
on the engineers to keep them posted as to how the caisson is going, 
for the superintendent who tries to do this for himself with a hand 
level or plumb line will be much astonished to find how far he is out. 
This is the reverse in building masonry, for a good mason, when 
started right, can generally be relied on to build a pier absolutely 
to lines. 

Fig. 2, Plate III, taken on October 13th, 1906, shows the nccdle- 
beams used to support the old building. As the underpinning has been 
described in the technical press, it will not be enlarged upon here. 

Fig. 3, Plate III, taken on October 18th, shows the work in 
full blast; Caisson No. 50 to the left is taking a rest while the 
concrete is hardening in the forms. Caisson No. 48-49, the second 
to be sealed in under air, is shown in the rear, with heavy cast-iron 
weights piled on top of the concrete. It also shows Caisson No. 30-43 
on the day when the compressed air was turned on, and Caisson 
No. 31-32, on which the lock is being placed. As has been stated, some 
1 200 tons of cast-iron weights and pig iron were in use on this job 
at one time, and it can be readily understood what an expensive item 
it was. The great advantage of the cast-iron blocks, weighing If tons 
each, is the saving in time in handling. The only disadvantage is the 
necessity of using the derrick when it may be wanted for other 
purposes. 

Fig. 1, Plate VI, gives a good idea of one of the serious diffi- 
culties which caisson men experience in city work, namely, lack of 
space; in fact, this is shown in nearly all the other plates. Here was 
a lock resting on the dumping platform, piles of temporary forms on 
the roadway, etc., and yet the contractors were obliged to be con- 
tinually hauling plant, etc., back to the yard, to be returned when 
needed, perhaps in a few days. 

Fig. 4, Plate III, taken on November 12th, 1906, shows whoro 
part of the dumping platform has been cut away for Caisson No. 28-34, 
the thirteenth to be sunk. It also shows 12 by 12-in. timbers piled up 
under the diimp and a section of shafting in the corner, with forms 
everywhere — on top of the shanties, loaning against the wall, etc. 

Fig. 2, Plate IV, taken on November 20th. shows a similar condi- 
tion, the only space left clear being room for a team to drive from 



14 PNEUMATIC FOUNDATIONS 

Broadway, under the derrick, where it could be turned around l)y a 
good driver. 

Fig. 3, Plate IV, a bird's-eye view, makes the unavoidable jumble 
look even worse. 

Fig. 2, Plate VI, shows a caisson which has just been lifted oft" 
the truck, and gives an excellent idea of its construction. The bolts 
shown extend down to the cutting edge, and are 1 in. in diameter and 
about 3 ft. apart; the temporary 2-in. plank roof, on which the men 
are standing, was removed about 48 hoi;rs after 2 ft. of concrete had 
been placed upon it. Fig. 2, Plate VT, also shows the special provision 
made for "hooking on" to the caisson for hoisting, which is much 
neater and better than the old way — still often seen — of wrapping 
around the caisson a rope or chain which was always likely to slip 
and injure it. 

The cutting edge is a 6 by 4 by ^-in. angle with the 6-in. leg 
horizontal; on top of the cutting edge is a horizontal course of 8 by 
12-in. (12 in. vertical), then four courses of 6 by 12-in., on top of 
which is a 10 by 12-in. course. Tlie longer caissons, of course, have 
struts and ties in the air-chamber. 

Fig. 4, Plate IV, taken on January 7th, 1907, gives Broadway 
a rather unnaturally deserted appearance, but gives a good end view 
of the combination collapsible shafts, and under them some bottom- 
dump buckets. 

Fig. 1, Plate V, taken on the same day, shows the cofi"er-dam 
on top of the forms of a caisson, but the writer can vouch for the 
fact that no caisson on this work was so much out of plumb as this 
picture would indicate; which goes to prove that the camera is some- 
times quite a liar. 

Fig. 2, Plate V, taken on January 21st, 1907, shows the four- 
masted derrick nearly all removed, the 28th caisson having been com- 
pleted on January ISth, leaving only two to be done, and, as these were 
to be directly under the derrick, they could not be placed until the 
derrick and its platforms had been removed, so it was January 23d 
and 24th, respectively, before these two caissons were located, and the 
13th and 19th of Fel)ruary before thoy were completed. Thus tho 
work was done in 154 days from the time of signing the contract — 
the 110 days of the contract being understood to cover tho time to be 
taken if the sinking stopped at Flevation 70; but all the caissons 



PLATE VII. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 





I'NEUMATIC FOUNDATIONS 15 

went below this depth, many below 90 ft., and the additional 20 ft. of 
sinking consumed more time than the upper 70 ft. 

Fig. 3, Plate V, was taken on January 2;3d, 1907, just before 
the 29th caisson reached the site. Caisson No. 3-9 (the 28th finished), 
though completed, still shows the lock and cast-iron weights in place. 
This view was taken looking toward Broadway, whereas the previous 
views were taken looking from Broadway. 

Fig. 1, Plate VII, taken on January 29th, shows an additional 
derrick, erected on three high trestles, just for handling these last 
two caissons. 

Fig. 3, Plate VI, shows the needle-beams and screw-jacks used 
in underpinning the old building; 40 and 6()-toii sprcw-jacks were 
used. 

Fig. 2, Plate VII, of February 16th, 1907, given a fair idea of 
the constant changes in the platforms, etc., as the work progressed 
from day to day. 

Fig. 3, Plate VII, taken on February 16th, 1907, looking toward 
Broadway, shows the last caisson three days before the air was taken 
off. It also shows how rapidly the place was being cleaned up in order 
to set the bases, finish the concreting, etc. 

One month later, that is, on March 16th, 1907, Fig. 4, Plate VII, 
shows that all the derricks of The Foundation Company have been 
removed, and that a big guy derrick for the iron erectors — Messrs. 
Milliken Brothers — has been put in place; the anchor-bolts, bases, etc., 
are everywhere in evidence. 

Fig. 1, Plate VIII, of ]\rarch 25th, shows two of these guy 
derricks in place, and Fig.' 2, Plate VIII, and Fig. 4, Plate VI, 
of April 5th, show a very different scene, the erection of the steel- 
work being well under way. These two views show very plainly how 
the columns were anchored to the caissons, the detail of the anchors 
being shown on Fig. 11. Mr. O. F. Semsch designed these with sec- 
tions decreasing from the top down to the bottom, at 60 ft. below the 
curb, the idea being to save the weight of the anchors by counting 
on the adhesion of the concrete to the steel bars, iising 50 lb. per 
sq. in. for adhesion, as allowed by the New York Building Department. 

Thus, for the bottom 10 ft., Mr. Semsch used one bar varying from 
6 by li in. to 6 by 31 in.; for the next lOi ft, above, he used two 
bars varying from 6 by 1^ in. to 6 by 2 J in., and coupled to the lower 



IG I'NEUMATIC FOUNDATIONS 

bar by pins 6i in. in diameter; for the next 9i ft, or from 40 ft. to 
30^ ft. below the curb, he used three bars varying from 6 by 1 in. to 
6 by 21 in. ; then four bars to 22 ft. below the curb, these running 
from 6 by 1 in. to 6 by If in. At this point he had a saddle with a 
7i-in. pin which connected the four flat bars with four round rods, 
which were 2| in. in diameter where the uplift was 270 tons, and 
u§ in. in diameter where it was 480 tons. These are the round rods 
projecting above the column bases as shown in the plates. They were 
ordered 2 ft. longer than the calculated length so as to allow for any 
variation in the depth of the caisson, etc. 

In three of the columns (Nos. 11, 16, and 22), the base was 8^ ft. 
lower, so the four top flat bars were omitted, the saddle being placed 
on top of the three flat bars. 

The columns anchored were interior columns, ten in all, viz., Nos. 
8, 11, 15, 16, 21, 22, 26, 27, 28, and 29. 

Table 1, showing loads on foundations, is made up from figures of 
Messrs. Boiler and Hodge. 

By the courtesy of The Foundation Company, the writer was 
allowed access to all records of construction, and, therefore, has 
been able to give, in Table 2, the complete record of each individual 
caisson. This table shows the number of caissons on the site and the 
number under air at any one time. It also shows the difference in 
time consumed in going through quicksand and hardpan ; for instance, 
in Caisson No. 17-18, the tenth to be sunk, 50^ ft. of sand were 
penetrated in 62 hours, while it took 91 hours to go through 20 ft. of 
hardpan. The writer has seen a caisson penetrate only 6 in. in 24 
hours, where boulders, etc., were encountered. 

It will be noticed that the total number of hours under compressed 
air is very much greater than the total number of hours of actual 
excavation and concreting. This was due to the fact that forms were 
used above the deck instead of coffer-dams, requiring the excavation 
to stop until the concrete was hard enough to allow the forms to be 
taken off safely. 

Where coffer-dams are vised, the sinking can be kept up con- 
tinuously, even if the concrete is being placed below the ground level, 
which, of course, is impossible with forms which have to be removed. 
Again, where the penetration is thus suspended temporarily, the quick- 



PLATE VIII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXlll, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 




Fig. ].— March ~5th, 11107, Two Guy Derricks Ready to Erect Steelwork. 




Fig. 3.— April 5th, 1907, Erection of Steelwork Well U.nder Way. 



1'-m:ijai.\tic foundations 







'i'M 


- 


':C:: 




O.'.O 




v"0'.' 




B.-o 




■■.?•••; 


— . 


oV-6 





&\-jJ- 




-.o-'. 




i!-;-|)- 




:-.o;; 




P^fo. 




vM 





>!V/.P;-''?.-.| 
'9':"6;"-.{o,';'vV; 

•■■.".? ^'.o'-yo-.Vj ■ 
■.•0:.• 



O'.V.-.V.O-'.'- 



•■■.■.••■■. ■.■5:-.-9.vo"-Ci,.,.„ . 
■O.-O ■■•.■.•-•-",■ .■"/^- •■-■;.■..■-'.■ O.-.o 



'■.P-.Va-: 



■oc-o'.'- 



..J/6:--K-'-.'.'o 



■.•o.::i';'.?-.:o-.-. <)■".■ 






■ ■.•••o.'.'.o-. 



1 r:i* 




IS 



I'MiUMATlC FOUNDATIONS 



TABLE 1. — Loads on Column Bases, in Tons. 

15 tons per sq. ft. allowed on caissons at top; no allowance for wind 

on caissons where wind load does not amount to 50% 

or more of the live and dead load. 



Column 
number. 


Dead load. 


60^ 
Live load. 

56.4 


Combined dead 
and live loads 


Wind loads. 


1 


756.7 


813.1 


747.1 


2 


614.3 


98.5 


712.8 


747.1 


3 


447.5 


108.9 


556.4 


373.5 


4 


447.5 


108.9 


556.4 


375.5 


5 


786.2 


88.4 


874.6 


747.1 


6 


884.2 


42.6 


926.8 


747.1 


7 


763.0 


79.9 


842.9 


747.1 


8 


289.2 


131.6 


420.8 


747.1 Anchor. 


9 


284.4 


129.6 


414.0 


373.5 


10 


284.4 


129.6 


414.0 


373.5 


11 


289.2 


1.31.6 


420.8 


747.1 Anchor. 


12 


763.0 


79.9 


842.9 


747.1 


13 


451.1 


94.9 


546.0 


373.5 


14 


284.4 


129.6 


414.0 


373.5 


15 


327 4 


147.4 


474.8 


747.1 Anchor. 


16 


327.4 


147.4 


474.8 


747.1 Anchor. 


17 


284.4 


129.6 


414.0 


373.5 


18 


445.4 


92.0 


537.4 


373.5 


19 


528.7 


92.6 


631.8 


373.5 


20 


284.4 


129.fi 


414.0 


373.5 


21 


327.4 


147.4 


474.8 


747.1 Anchor. 


22 


327.4 


147.4 


474.8 


747.1 Anchor. 


23 


284.4 


129.0 


414.0 


373.5 


24 


454.7 


94.3 


.549.0 


373.5 


25 


602.5 


91.8 


694.3 


747.1 


26 


289.2 


131.6 


420.8 


747.1 Anchor. 


27 


284.4 


129.6 


414.0 


747.1 Anchor. 


28 


284.4 


129.6 


414.0 


747.1 Anchor. 


29 


289.2 


131.6 


430.8 


747.1 Anchor. 


30 


602.5 


91.8 


694.3 


747.1 


31 


596.4 


68.3 


664.7 


650.3 


32 


614.3 


98.5 


712.8 


747.1 


33 


447.5 


108.9 


556.4 


373.5 


34 


447.5 


10H.9 


.556.4 


373.5 


35 


614.3 


98.5 


712.8 


747.1 


36 


596.4 


68.3 


664.7 


747.1 


37 


73.2 


46.4 


119.5 




38 


133.3 


85.1 


218.4 




39 


191.1 


121.8 


312.9 




40 


131.8 


83.9 


215.7 




41 


272.2 


25.7 


297.9 




42 


214.4 


18.8 


233.2 




43 


214.4 


18.8 


333.2 




44 


142. K 


9.4 


152.2 




45 


120.3 


27.2 


147.5 




46 


113.6 


43.2 


186.8 




47 


143.(i 


43.2 


18(i.H 




48 


143.ti 


43.2 


186.8 




49 


120.3 


27.2 


147.5 




50 


275.9 


36.9 


312.8 




51 


350.0 


52.0 


402.0 




52 


242.4 


37.9 


280.3 




53 


242.4 


37.9 


2H0.3 




54 


328.5 


29.8 


358.3 





PLATE IX. 

TRANS. AM. 30C. CIV. ENGRS. 

VOL. LXIII, No. t099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 



TABLE 2. — Recokd of Sinking Caissons foe the Sinseb Building, 1906-1907. 



NiTMBBR OF Caisson. 


50 


48-49 


40-41 


30-43 


45^6 


31-32 


5-6 


39 


19-20 


17-18 


54 


30-42 


28-34 


53-53 


7-8 


Order of sinking. 


] 


3 


8 


4 


5 


6 


7 


8 


9 


10 


11 


12 


13 


14 


15 


Size of caisson. 


r O" X r 2" 


5- 6" X 16- 4" 


7- 2" X 11- 10" 


e- 2" X 16- 4" 


6-6- X 16' 4- 


7- 8" X IS- 4- 


7- 8- X 21- 2" 


6- 6" diam. 


6- 6- X 17' 10- 


5- 6- X 16 8 " 


7' 0" X 8- 2" 


6- 2" X 16- 4 - 


6- 2" X 18- 8- 


5- 6" X 13' 8' 


7- 8-- X 18- 4" 




5,30 p. M. 
Sept. 29 


9.00 p. M. 
Oct. 11 


4.80 p. M. 
Oct. 5 


1.45 p. M. 
Oct.O 


4.80 p. a. 
Oct. 16 


9.00 p. M. 
Oct. 6 


0.30 p. M. 
Oct. 12 


7.00 p. M. 
Oct. 26 


9.30 p. H. 
Oct. 5 


6.45 p. M. 
Oct. 9 


5.30 p, M. 
Nov. 9 


4.30 p. M. 
Oct. 8 


11.85 A. M. 
Nov. 10 


7.20 p. M. 
Nov. 13 


8.00 p. M. 
Oct. 13 




Compressed air on -j gjj?. 


10.30 A. M. 

Oct. 15 


1.20 p. M. 
Oct. 17 


1.30 A. M. 

Oct. 16 


6.00 P.M. 
Oct. 18 


4.30 a. u. 
Oct. 27 


1.00 a. m. 
Oct. 26 


9.30 A. M. 
Oct. 29 


6.00 p. M. 
Nov. 11 


9.30 p. M. 
Oct. 31 


9.0O p. M. 
Oct. 30 


4.00 p. M. 
Nov. 16 


12.30 p. M. 
Nov. 13 


3.00 A. H. 

Nov. 18 


7.00 A. M. 

Nov. 30 


5.00 p. M. 
Nov. 18 


Caisson reached hardpan j Date! 


2 a.m. 
Oct. 21 


6.00 p. M. 
Oct. 21 


10.00 A. M. 

Oct. 83 


2.00 p. M. 
Oct. 27 


9.30 A. M. 
Oct. 31 


8.00 A. u. 
Nov. 4 


10.8-) A. M. 
Nov. 6 


11.00 A. M. 

Nov. 13 


10.00 p. M. 
Nov. 15 


8.30 a. M 
Oct. 31 


7.00 p. M. 
Nov. 21 


6.00 A. M. 

Nov. 38 


11.00 p.m. 
Nov. 27 


8.00 p. M. 
Nov. 80 


1.30 a. m. 
Nov. 37 




6.30 A, M. 
Oct. 28 


9.00 p. M. 
Oct. 24 


9.30 p. M. 
Oct. 24 


8.00 p. M. 
Oct. 29 


9.00 A. M. 
Nov. 2 


6.00 a.m. 
Nov. 10 


7.00 A. M. 
Nov. 11 


11.00 a.m. 
Nov. 14 


2.80 p. M. 
Nov. 10 


7.30 p. M. 
Nov. 19 


11.00 a.m. 
Nov. 28 


11.80 A.M. 

Nov. 37 


2.0O p. H. 
Dec. I 


8.00 p. M. 
Dec. 2 


3.30 p. H. 
Dec. 8 




Concreting begun in air-chamber -1 ^^te" 


Oct. 82 


10. IB p. M. 
Oct. 84 


10.00 A. M. 

Oct. 25 


8.80 p. M. 
Oct. 29 


10.00 a.m. 
Nov. 2 


7.00 A. M. 
Nov. 10 


9.30 A. M. 
Nov. 13 


18.80 p. M. 
Nov. 14 


3.00 p. M. 
Nov. 19 


8.45 p. M. 
Nov. 19 


11-30 A M. 

Nov. 83 


12.30 p. M. 
Nov. 37 


3.30 p. M. 
Dec. 1 


8.30 A. M. 
Deo. 2 


3.00 p. M. 

Deo. 3 




4 A.M. 
Oct 23 


r.OO A. M. 
Oct. 26 


12.30 p. M. 
Oct. 86 


2.00 p. M. 
Oct. 30 


10.00 A. M. 
Nov. 3 


8.00 A. M. 
Nov. 11 


4.00 p. M. 
Nov. 14 


6.00 a.m. 
Nov. 15 


4.00 p. M. 
Nov. 21 


8.00 p. M. 
Nov. 21 


11.59 p.m. 
Nov. 23 


10.30 A. M. 

Nov. 30 


3.80 p. M. 
Dec. 3 


4.00 A. M. 

Dec 4 


1 30 p. M. 
Dec. 5 




Number of hours under compressed air 


185« 


209H 


261 


286 


178)^ 


891 


390 J« 


84 


498^ 


637 


176 


382 


378^4 


883 


S48H 




Hardpan 


Hocli 


Hardpan 


Hardpan 


Hardpan 


Kocli 


Rocit 


Hardpan 


Rocli 


Hocl! 


Hardpan 


Rock 


Bock 


Hardpan 


Bock 






nitehing..., iS''^*- 


5 
8 


5 
3 


5 
3 


6 
3 


3 


5 
8 


5 
S 


5 
8 


6 
8 


6 
8 


3 


6 
3 


5 
8 


5 
8 


5 






Excavation in quicksand ] Horns 


40 
70 


61 
67 


47 
77 


49 
02 


44 
55 


44 
70 


5"H 

82 


46 
41 


48 
66 


50« 
62 


t^ 


^»« 


5014 
88 


46 
57 


42« 

77 


Excavation in hardpan -j g^'J^^ 


10 
2? 


13 
69 


n 

35 


7W 
81 


11!4 
36 


2.'i 
142 


111 


82 
24 


16!^ 
88 


20 
91 


13 
40 


21 
103 


3^ 
87 


14 
48 


21 
133 




SI- 
IS- e- 

84- 

74- 6" 

m- 6" 



99 

101 


15' 
28- 1" 
71- 4" 
84- 4- 

73- 

244 
208 


IB- 
18- 8" 

67- 

77- U" 

68- 6" 

4 

197 

183 


IB- 
18- 6- 

69- 

70- 7" 

71' 4- 

4 

229 

213 


18- 
23- 1" 

67- 

78- 6" 

69- 0- 



213 

183 


18- 
18- 8" 

67- 

92- 

68- 6- 



385 

378 


15- 
27- 2 - 
70- 6- 
89- 11" 
72 5-- 

14 
449 
377 


18- 
23- 1" 

69 
77-5" 

71- 


78 

72 


20- 
18' a- 

08- 
84- 7- 
69' r - 


334 

287 


15' 
27- 2" 
70- 0" 
90-6" 
70- 6" 
14 

356 

315 


IB- 
IS' m" 

66- 5-- 

78-5" 

68-5- 

4 

135 

134 


LI- 
IS- »%■■ 
09 6 
90- 6- 
09- 9- 
4 

283 

365 


15' 
18' 8" 
70- 3" 
86' 9- 
70- 6 " 
7 
273 
359 


li' 

17' 6 " 

64' 11" 

78' 11" 

67- 11 - 

4 

177 

171 


30- 


" top of concrete 

*' " " " " " liardpan 

" " " '■ bottom of excavation. 


18' 8- 
67' low- 
88- IOjI" 
69- 4)i" 


Number of feet of coffer-dam above concrete 

" cubic yards of excavation 

" " concrete 


4 
368 
.363 



PLATE X. 

TRANS. AM. 80C. CIV. ENQRS. 

VOL. LVIII, No. 1099. 

THOMSON ON 

PNEUMATIC FOUNDATIONS. 



TABLE 2. — Record of Sinking Caissons for the Singer Building, 1906-1907 (Continued). 



NaMBBR OP Caisson. 


■H-H 


18-14 


47 


11-12 


61 


1-2 


27-33 


29-:i5 


37-38 


4-10 


28 


26-26 


3-9 


16-22 


15-21 


Order of sinking. 


10 


17 


18 


19 


20 


21 


22 


28 


24 


26 


26 


27 


28 


29 


30 


Size of caisftOD. 


5' 6" X 16- 4" 


5- 6" X 16- lO" 


8' 0" diam. 


r'8"x 18' 8' 


5' d" X- 11' 8" 


7- 8 " X 19' 4" 


6' 2" X lO- 8' 


7' 8" X 17- 8" 


5' 6" X 16' 2" 


5' 6 " X 16' 8' 


7' 0" X 7' 0" 


17' 8" X 17' 4" 


5' 6" X 16' 8" 


7' " X 17' " 


7' 0' X 17' 0" 


Caissou arriveJ on lot ] p^J^ 


4.01) p. M. 
Oct. 9 


3.00 p. M 
Oct. 23 


7.00 p. M. 
Oct. 26 


Noon 
Nov. 6 


1 00 p. «. 
Nov. 28 


11.46 A. H. 

Oct. 31 


.8.00 p. H. 
Dec. 6 


4.00 p.m. 
Nov. ai 


3.20 p. H. 
Dec. 5 


2.00 p. u. 
Dec. 19 


8.00 p. H. 
Deo. 24 


8.46 p. M. 
Dec. 16 


7.00 p. M. 
Dec. 19 


6.00 p. M. 
Jan. 23 


Noon 
Jan. M 


Compressed air on ] Date. 


8.00 A. H. 
Nov. 19 


5.00 A. H. 

Nov. 21 


8.;«P.B. 
Dec. 9 


8.00 A. H. 

Bee. 3 


6.80 p. B. 
Dec. 7 


Noon 
Deo. 3 


2.30 A. H. 
Dec. 14 


10 A. M. 

Dec. 11 


2 p.m. 
Dec. 15 


3.00 p. M. 
Dec. 29 


2.00 A. M. 

Jan. 4 


12.30 p. M. 
Jan. 2 


8.00 p. M. 
Dec. 30 


Noon 
Feb. 5 


12.30 A. M. 
Keb. 4 




8.1111 i. 11. 
Nov. 25 


11.00 A. M. 

Dec. 6 


8.00 A. M. 
Dec. 11 


4.00 A. M. 
Dec. 7. 


9.30 A. H. 
Dec. 15 


5.30 a. u. 
Dec. 20 


10.30 p.m. 
Dec. 33 


6 p.m. 
Dec. 21 


4.00 A. M. 

Jan. 4 


4 00 p. M. 
Jan. 6 


7.00 A. M. 

Jan. 10 


11.30 A.M. 

Jan. 10 


2.00 p. H. 
Jan. 12 


8.00 p. M. 
Feb. 8 


10.00 p. M. 
Feb. 14 






10. 4B A. M. 
Dec. 8 


10.00 p. u. 
Dec. 9 


3.00 p. u. 
Dec. 12 


7.00 A. H. 

Dec. 13 


10.00 A. u. 
Dec, 17 


11.00 p. M. 
Dec. 26 


2.30 p. H. 
Dec. 30 


7 A.M. 
Jan. 6 


4.00 p. M. 
Jan. 6 


2.46 p M. 
Jan 9 


7.0U A. M. 

Jan. 14 


7.00 A. M. 

Jan 16 


7.30 p. M. 
Jan. 15 


11.30 p. M. 
Feb. 11 


8.30 A. M. 
Feb. 18 




Concretinff begun id air-chamber — j p^te 


1I.:»A.M. 
Dec. 8 


lo.ao p. M. 
Dec. 9 


3.30 p. M. 
Dec. 12 


7.30 A. M. 
Dec. 13 


10.80 A. M. 
Dec. 17 


11.30 p.m. 
Dec. 26 


3.00 p. u. 
Deo. 30 


7.30 A.M. 
Jan. 6 


4.80 p. H. 
Jan. 6 


3.20 p. M. 
Jan. 9 


7.80 A. M. 
Jan. 14 


7.30 A. M. 
Jan. 16 


8.30 p. M. 
Jan. 15 


11.46 p.m. 
Feb. 11 


9,00 a.m. 
Feb. 18 


Compressed air taken off \ g^J^J* 


5.00 A. M. 

Dec. 10 


1.30 p. u. 
Deo. 11 


Noon 
Dec. 13 


1.00 a. w. 
Dec. 15 


5.30 A. M. 
Dec. 19 


8.80 A. H. 
Dec. 29 


12.30 p. M. 
Jan. 3 


12.30 A. M. 
Jan. 9 


11.30 p. M. 
Jan. 7 


6.30 a.m. 
Jan. 11 


10.30 A.M. 

Jan. 15 


10.80 A.M. 

Jan. 18 


9.80 A. M. 
Jan. 18 


4.30 P.M. 
Feb. 13 


10.15 p. M. 
Feb. 19. 


Number of hours under compressed air 


501 


488^ 


«ni 


281 


274^4 


ma 


490 


686)4 


661H 


303!^ 


ma 


382 


446)i 


19614 


381l< 




Rock 


Rock 


Hardpan 


Rock 


Hardpan 


Rock 


Rock 


Rock 


Hardpan 


Rock 


Rock 


Rock 


Rock 


Rock 








■"'""tag I&^u'^ 


5 

8 


5 
3 


6 
2 


5 
3 


5 
3 


5 
8 


6 
8 


5 
3 


6 
8 


6 
3 


5 
3 


5 

3 


6 
3 


5 
3 


5 
8 


Excavation In quicksand { Hmire 


53 

58 


42 
56 


62 
35 


49 
79 


49 
52 


48 
75 


60 

77 


60 
96 


43 

53 


y,^ 


49 
48 


41 
64 


51 

68 


60 
74 


60« 
67 


Excavation in hardpan ) g««;. 


18.1 
98 


19 
83 


10 
81 


22 
140 


14^4 
46 


21 
18U 


20 
119 


30 
1'22 


12 
60 


18 
71 


2m 

96 


26 
189 


18 
67 


19 
78 


17 
84 


Distance from curb to groucd 

" top of concrete 

" ' " iiardpan. 

" " *' " liottom of excavation. 


15' 
S7' 2- 

n- 

00- 8" 

72- 9" 

14 

267 

209 


20' 

W 8" 

ar 

88' 1" 
68' 1" 
4 
220 
228 


16' 
28' I" 

72' 

82' 
73' 6" 

12 

82 

82 


15' 
27' 2" 

69' 

90' 8" 

70' 8" 

18 

390 

327 


15- 

19' 5" 

69' 
88' 6" 
70' 6" 
4 
168 
15a 


20' 
18' 8" 

67' 

g7' 8" 

68' 8" 

4 

371 

877 


15' 

18' 8" 

70' 

m 

68' 

10 
285 
Wl 


15' 
18' 8" 
m: 2" 
90' 2" 

72' 

10 

srr 

358 


20' 
18' 8" 
67' 9" 
79' 9" 


16- 
27' 3" 
71' 6" 
89' 7 " 


15' 
27' 3" 

69' 

90' 6" 

70' 3" 

14 

137 


30' 

18' 8" 
66' 
91' U " 
72' B" 

354 


20' 
23' 1" 

71- 

89' 
78' 3" 

14 

261 

234 


16' 
OT'2" 
69' 9" 
89' 9" 
70' 9" 
14 
329 
276 


15' 
18' 8" 
70' 7" 
87' r 


Numljei- of feet of coffer-dam above concrete 
" cubic yards of excavation 


4 10 
l»r 363 


4 
320 






1 









PNEUMATIC FOUNDATIONS 19 

sand is apt to pack against the caisson, greatly increasing the friction 
and requiring much more pig iron to overcome it. 

This form of construction shows to the greatest advantage where 
it is possible to place all the concrete on the caisson before the sinking 
commences, or where the total depth of caisson and coffer-dam will 
not be more than 30 ft. 

As soon as a caisson was placed and had some weight on it, it was 
"ditched," that is, sunk — generally about 5 ft. — without air, so as to 
make it safe to add the locks and more concrete and to make the 
bracing against overturning easier. 

Instead of giving an average case or extreme cases, both of which 
are usually misleading, the history of every caisson is given in Table 2. 

The writer has endeavored to confine this description to what 
was out of the ordinary, and the unique feature of the foundations of 
the Singer Building was certainly the undermining of Caisson No. 
30-43. In tabulating No. 30-43, the results are those obtained when 
it was originally sunk, and do not include the tunneling. 

Of course, a contractor always wants to stop pumping compressed 
air into a caisson as soon as he can do so with safety, after it has 
been filled with concrete; but it would be much better to keep up the 
air pressure for at least from 12 to 24 hours after the concreting is 
completed. The writer has repeatedly found that if this were done 
he could make the concrete in the air-chamber water-tight, whereas, if 
it were not done, the water would rapidly force its way thiough the 
concrete to the top, and nobody will claim that it is advisable to have 
water flowing through concrete before it has set. 

It has been stated that the sample of concrete taken from the 
bottom of Caisson No. 30-43, when it was undermined, was of the 
hardest, showing that the concrete had set sufficiently before the air 
was taken off. 

In Table 2, the caissons are designated by the number of the 
columns resting on them, as originally laid out on the column plan. 
Fig. 2. 

The entire work was done under the personal direction of Mr. C. P. 
Coleman, the Executive Officer of the Singer Manufacturing Company, 
and Mr. Ernest Flagg, his Architect. Mr. 0. F. Semsch was Chief 
Engineer for the Architect, and Mr. IT. J. Howells had charge of the 
inspection for him. A. P. Boiler and IT. W, Ilodge, ^fcmbers, Am, 



20 PNEUMATIC FOUNDATIONS 

Soc. C. E., were the Consulting Engineers for the building, the Con- 
tractors being The Foundation Company, of which Mr. Franklin Rem- 
ington is President, and D. E. Moraii, M. Am. Soc. C. E., E. S. 
Jarrett, Assoc. M. Am. Soc. C. E., and L. L. Brown, M. Am. Soc. C. E., 
are members. Mr. Alexander Allaire was Superintendent in charge for 
the Contractors, and the writer was Consulting Engineer for the 
caisson work for the owners. 

He desires to thank the owners, architects, engineers, and con- 
tractors for their kindness in furnishing information, records, etc., 
for this paper. It is unnecessary to state that their work on the 
building was well done. 



DISCUSSION ON PNEUMATIC FOUNDATIONS 21 

DISCTJSSIOlSr. 



O. F. Semsch, Esq. — Mr. Thomson states that some of the columns Mr. 
of the Singer Tower were anchored to the footings in order to provide 
for the uplift caused by the wind pressure on the building. A brief 
description of the wind bracing, and more particularly of the method 
of calculating the load due to wind pressure on the footings, may prove 
of interest. It was assumed that the tower would be exposed to a wind 
pressure of 30 lb. per sq. ft. from the top to the roof of the 14-story 
main building, the lower portion being sheltered from wind on all 
sides by the surrounding buildings. 

The wind moment equals 23.7% of the moment of stability, both 
moments being taken about the base of the tower. As the Building 
Code of New York City specifies that the wind moment of any 
structure shall not exceed 75% of its moment of stability, the design 
is well within the limits of the law. 

There are eleven sets of braces to resist the wind pressure in a 
northerly and southerly direction, and ten braces in an easterly and 
westerly direction. These extend up to the thirty-third story; below 
the fourteenth story there are two additional braces in each direction. 
Each wind brace may be said to consist of a vertical truss 12 ft. wide 
and about 500 ft. high, formed by two lines of columns with cross- 
latticing between them. 

On account of the unusual architectural treatment of the tower 
fagades, it was impossible to install a system of bracing extending 
across the entire width of each side, which, of course, would have 
been the most natural thing to do. Each front of the tower consists 
of five equal bays, the three middle ones being combined into what 
is practically one huge window about 36 ft. wide. This window pro- 
jects several feet beyond the wall line; it was impossible, therefore, 
to run any bracing across it. Accordingly, four braces were placed 
in each corner of the tower, forming really four small towers, each 
12 ft. square, and the others were grouped around the elevator shafts 
in the center of the building. 

Each brace was calculated so that it could stand alone and resist 
its share of the wind pressure. Thus, if ten of these braces were to 
be set side by side, with a plate, about 60 ft. wide by 500 ft. high, placed 
against them, they would be capable of resisting safely a pressure of 
30 lb. per sq. ft. on that plate. 

This method of calculating the wind load, according to what the 
speaker terms the "brace" system, to distinguish it from another 
method to be mentioned later, resulted in concentrating that load, as 
far as the foundations were concerned, under the sixteen columns at 
the four corners of the tower and under the elevator-shaft columns 
in the center. The designers, however, were sure that in reality the 



32 DISCUSSION ON PNEUMATIC FOUNDATIONS 

entire wind load would not be transmitted to the footings by these 
columns alone, but that some pressure would be carried down by each 
column of the tower, whether or not it formed part of the bracing. 
In order to get some idea of what this load for each column might 
be, the tower was regarded as a monolith, and a calculation was made 
of the distance the resultant of the wind load and the total weight of 
the tower would fall outside of the center at its base. This distance 
was found to be 4 ft. From this was ascertained the ratio according 
to which the average pressure under each column would be increased 
on account of this eccentricity. In this way the amount of wind load 
that each column would transmit to the footings was found, according 
to what might be termed the "monolith" system. 

There were then two sets of calculations, and finally, in accordance 
with the suggestion of Mr. Ernest Flagg, the architect, the average 
between the loads obtained by each method was taken as the wind load 
on the footings, for it seemed reasonable that the correct solution of 
the problem lay somewhere between the two. In designing the anchors, 
however, the full uplift developed by the braces regarded as standing 
alone was taken. 

Messrs. Boiler and Hodge, who made a separate calculation, as- 
sumed that two-thirds of the entire wind load would be transmitted 
to the footings by the columns forming part of the wind braces, and 
the remaining third by the columns outside of the wind braces. 
Wherever the wind load at the base of a column amounted to less 
than 50% of the combined dead and live loads, they disregarded it 
altogether — according to the practice followed in bridge designing. 

The wind loads given by Mr. Thomson in Table 1 were really 
reduced by one-third before being applied to the footings, because the 
Building Code of New York City allows an increase of 50% in stresses 
when calculating for wind load. 

According to the architect's method of calculating, the total load 
on the tower footings proved to be, in round numbers, 30 000 tons. 
This, divided by the total caisson area, 2 794 sq. ft., gives an average 
pressure of 10.7 tons per sq. ft. The load is not evenly distributed, 
however, but varies from 15 tons per sq. ft. in some places to 5 tons 
in others. Messrs. Boiler and Hodge allowed 15 tons per sq. ft. for 
live and dead loads, and 22^ tons per sq. ft. for live, dead, and wind 
loads. According to either method, the tower is amply safe, and its 
rigidity and solidity are really remarkable. 

As mentioned by Mr. Thomson, several of the caissons were out 
of plumb a few inches, and one was out about 1 ft. Considering the 
depth, 90 ft., this was not to be wondered at. However, when the 
ovmer heard of it, he accepted the situation as soon as the Foundation 
Company had agreed to put a 12-in. bed of concrete over the entire 
area, and around the tops of the caissons, tying them all together. 



PLATE XI. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1099. 

SEMSCH ON 

PNEUMATIC FOUNDATIONS. 




The Si.vger Building. New York City. Dl-ring Erection. 



biscussio^: ON pneumatic foundations 



23 



Mr. Semsch. 



I\ 



Average width 
eiposfd to Wind 

= d'o" 



2 1-^^350 X (13 X 20 lb. 
-1 



= iJiOUOlb. 
(On accountjof 50^ 
incFeaee in streas 
allowed by 

Building Code.) 



E1.20l'6" 



Wind moment about 
= «S0001Ux380 d 
168 000 000 it.-lb 




DIAGRAM OF INCREASED PRESSURES 
ON COLUMN FOOTINGS OF SINGER BUILDING DUE TO 
WIND; REGARDING THE TOWER 
AS A MONOLITH 



Windi 
221 Tons 



;o'o" '' 

Area of Base =: TO'Ox 72'o° 



To tret the Increase due to Wind, multiply the Load on the 
inner four Columns by l.OJ ; on the next square 
of 12 Columns by MS; on the outer 20 Columns by 
1.287. (Aicording to the Monolith Theory;. 




Average piessure = }^ *°" =3.05 tons per sq.ft. 
lO X i2 

Ratios: -:;^=IM; f^ = l.lS: -^-^=1.287. 
o.Go 3.0.J 



3.85 



Fig. 12, 



u 



biSCUSSIOX ON PNEUMATIC FOUNDATIONS 



Mr. Semsch, 



This precaution was really not necessary, because the surround- 
ing sand and hardpan held the caissons so securely in place that they 
could not have moved, but the owner very properly took the position 
that he was entitled to caissons which would be capable of standing 
up without any earth around them whatever. 

This method of solving the difficulty may be used as a precedent 
by engineers if their clients ever refuse to accept caissons on account 
of their being out of plumb. 



Mr. Stern. 




PLAN SHOWING LOCATION OF WIND BRACING 
IN SINGER TOWER 

Fig. 13. 

Eugene W. Stern^ M. Am. See. C. E. — Mr. Thomson has mentioned 
the inaccuracy of wash borings, and the speaker wishes to endorse this 
strongly. The results of these borings must be examined very care- 
fully and, if possible, compared with others taken in the neighborhood, 
and even after this is done the results are not always to be trusted. 
In one very serious case, wash borings were taken, which showed 
rock 30 ft. below the surface, and the foundations were designed 



DISCUSSION OX PNEUMATIC FOUNDATIONS 25 

ncc()r(liii<ily. On ?;iiil<iiig- the fitiUKhitioiis it was t'mmd that, instead of Mr. Stern, 
solid rock, the bottom consisted of broken stone filled in to a depth 
of 20 ft. on poor, soft bottom, the site having been originally a pond 
which was used as a convenient dumping ground for whatever was 
excavated in the neighborhood. Had this been known before the 
foundations were designed, many thousands of dollars would have been 
saved to the owners. 

The method of placing a contract for foundations, namely, a lump 
sum to a certain depth, and an extra price per cubic yard for every- 
thing below that depth, commends itself also as being fair to both 
owner and contractor. 

It would be interesting to know why it was decided to carry the 
foundation to rock after hardpan had been reached. Was not this 
considered sufficiently good foundation ? The hardpan excavated in 
the neighborhood of the Singer Building is practically of the con- 
sistency of concrete. The speaker has had a specimen exposed in his 
office for months without the slightest sign of disintegration. 

Mention is also made of the care required in sinking caissons 
in order to avoid tilting. This is, of course, very important, and too 
much care cannot be taken in this matter. It is money well spent. 
A dollar spent in avoiding misplacements will save hundreds later in 
trying to rectify them, and, in some cases, correction is impossible. 
It is necessary, not only that the work be started exactly right, but 
also that continual and careful observations be made with surveying 
instruments, in order that accuracy of position may be maintained. 

The methods of sinking pneumatic foundations have been carried 
to a great degree of refinement in New York City, and perhaps nowhere 
else in any country has so much work of this kind been done, and so 
much experience been gained. The Moran air-lock is undoubtedly the 
greatest improvement that has been made in many years in connection 
with the actual processes in pneumatic work. There are some things, 
however, which the speaker's experience suggests as being capable of 
improvement, namely, that working chambers, which are usually made 
of wood, might better be made of reinforced concrete; and that the 
shafts connecting air-locks with working chambers are usually made too 
small to permit a man to climb the ladder in safety and avoid the 
hoisting bucket. 

The speaker also agrees with Mr. Thomson that the process of 
sinking should be continuous, and that there should be no stoppages. 
If this is not done, the danger of the caisson being "hung up" is very 
great, and it is sometimes impossible to start the sinking again — even 
by loading it with cast-iron blocks — without blowing out the air; and 
this should never be done in building construction, as it is very likely 
to endanger the foundations of adjoining buildings. 

Continuous sinking is not po.ssible where the form method of 



26 DISCUSSION ON PNEUMATIC FOUNDATIONS 

Mr. Stern, placing Concrete is used I'ather than the co£Per-dam method, and, while 
the latter method is perhaps a little more expensive in first cost, it is 
very questionable whether it is not cheaper in the end, for it enables 
the work to be done more accurately and quickly, 
ir. Jarrett. Edwin S. Jarrett, Assoc. M. Am. Soc. C. E. — The speaker dis- 
agrees in one particular with ]\Ir. Stern. In so far as he criticizes 
adversely the practice, exemplified in the Singer Building foundation 
work, of moulding the concrete piers which surmount the caisson 
inside of temporary forms and of removing the forms before sinking 
the moulding piers, his views are not tenable. He states, in effect, 
that the long delays necessary at the various stages of the sinking 
operations to insure the setting of the concrete before it is safe to 
remove the forms and to continue the sinking have most injurious 
effects. The opportunity thus given for the increase of friction by the 
settling and compacting of the soil around the piers compels a recourse, 
as Mr. Stern views it, to excessive blowing in order to get the caissons 
to resume their downward movement. This blowing, in his opinion, 
very often throws the caissons out of plumb, and in general results 
in the loss of control of their movements. He favors a reversion to 
the old style of permanent coffer-dam forms because, in his belief, 
they allow a continuous sinking and thus insure, with reasonable care, 
complete control and better average results. 

It may be stated at once that the continuous sinking of caissons 
is a partial insurance against unskilful methods and careless handling. 
On the other hand, with proper care and the skill acquired by long 
experience, caissons may be held up almost indefinitely and sinking 
may be resumed at any time without any bad results, it being premised, 
of course, that all the obvious precautions shall have been taken. The 
specific evil dwelt upon by Mr. Stern — the sinking of caissons out 
of plumb — results only from careless work. The presence or absence 
of a permanent coffer-dam form has little to do with it. If a structure 
is carried plumb to a depth of, say, 20 ft., it is not easy to throw it 
out of level. Only such excessive blowing as would allow material to 
flow in under the cutting edge could accomplish it, and such blowing 
is reckless and is not allowed by a careful contractor. If the caisson 
is carrying ample weight, it can be taken down plumb, even after a 
prolonged stoppage in the sinking. It is concluded, therefore, that the 
use of removable forms does not necessarily entail the bad work which 
Mr. Stern considers has resulted from their wide adoption. 
• Without discussing this particular matter any further, it may 

be stated that a small economy and an increased efiiciency have resulted 
from this improvement in caisson work for buildings. 

The construction and methods described by the author have been 
brought out somewhat fully in the technical journals. Mr. Thomson, 
however, has dwelt on the novel feature of this particular work, namely. 



DISCUSSION ON PNEUMATIC FOUNDATIONS 37 

tunneling through the hardpan from caisson to caisson and under- Mr. Jairett. 
pinning to rock a completed pier more than 50 ft. high. Aside from 
the interest of this operation, it has a significance which might, under 
conceivable circumstances, be full of possibilities. The ease with 
which tunnels, conduits, and the like could be carried through this 
hardpan as compared with the great difficulty of work, either in the 
quicksand above or in the rock below, might be of no small conse- 
quence. As a rule papers of this kind make no mention of certain 
emergencies which arise in such operations. Accidents occur in the 
progress of foundation work, and require great skill and nerve in 
handling, particularly when the work is being carried on close to 
heavy buildings. In such locations all the quicksand released will 
probably come from under the adjoining heavy structures. 

It may be pertinent to mention an incident of this kind which 
occurred in the sinking of these foundations and shows the possi- 
bilities of damage, the great care which must be exercised, and the 
troubles that will arise in spite of that great care. 

The steam for operating the compressors on this particular work 
was supplied by the New York Steam Company, the pipes of which 
are laid in the streets in the lower part of the city. All steam for 
such work in Lower New York City is thus supplied, as there is no 
available space upon which to erect boilers. It is necessary, of course, 
to keep the compressors in continuous operation, because there are 
always two or three caissons on the way down, and the loss of air 
would cause an inrush of quicksand, which it is absolutely necessary 
to prevent. To do this, the air pressure must be kept up, and of 
course the steam must be steadily supplied. 

While this particular work was being done, the New York Steam 
Company's plant caught fire one night and, as the fire spread, boiler 
after boiler went out of commission. The air pressure began to 
decrease. There was one caisson about 15 or 20 ft. down, another 
more than half way down and a third was close to the hardpan. Had 
all the pressure been taken from the caissons without making pro- 
vision to keep out the quicksand, it is extremely likely that great 
damage would have resulted. Having ascertained that all the steam 
pressure would be withdrawn, quick action was necessary. The only 
course open was to flood the caissons. Streams of water were thrown 
into each, and the working chambers, together with the shafts con- 
necting them with the air-locks, were filled as rapidly as possible to a 
height above the water-level in the soil. There was thus obtained inside • 

the caisson a compensating pressure against the water in the soil, which 
eliminated the danger that quicksand might enter. 

After the air pressure was taken off, the caissons remained in the 
condition described, full of water, for from 24 to .36 hours. The com- 
pressors were then started, and air was applied to each caisson cau- 



28 DISCUSSION ON PNEUMATIC FOUNDATIONS 

Mr. Jarrett. tiously, some of the water being blown out and the remainder being 
forced back into the soil by the air pressure. When the water sub- 
sided it was found that in no caisson had quicksand entered the 
working chamber, but it was in substantially as good condition as 
when the air pressure was taken off. 

The flooding of caissons is resorted to quite often to drown out 
fires which in timber caissons are of not infrequent occurrence. The 
methods resorted to in order to meet siich emergencies are the natural 
ones, but very often, in the middle of the night, with no one around 
to think and act quickly, and with no one on the ground who has been 
through a similar experience, such situations are hazardous. 
Mr. Thomson. T. Kennard THOMSON, M. Am. Soc. C. E. (by letter). — The writer 
regrets that more members have not accepted his invitation to discuss 
this paper, or at least to ask questions about it. 

The Society is indebted to Mr. Semsch for his thorough discussion 
on the wind bracing of the Singer Tower. 

The writer has been asked whether full air pressure was required 
in hardpan. Generally, when the cutting edge has entered a foot or 
so into good hardpan and has been well plastered with clay, it is 
possible to carry a little less pressure than the depth at the bottom of 
the excavation would require, but, if the hardpan is poor, or if a vein 
of sand is encountered, the full theoretical head is necessary. 

Although some hardpan is very good, it is not all as good as con- 
crete, nor anything like it, as may be judged from the typical cross- 
sections. When the owners of the Singer Building found that there 
were soft streaks in the hardpan they decided to sink all the tower 
caissons to rock and not run any chances of undermining by some 
future tunnel from Jersey City or Brooklyn. This was a very wise 
precaution, especially when the narrow base is considered. 

Mr. Jarrett has mentioned the only proper thing to do when the 
air supply is cut off; the improper thing might also be mentioned: 
A few years ago the night superintendent on a certain foundation 
"lost his head" and pumped the water out of the air chamber and filled 
it up with sand, thus causing very serious settlement in an adjoining 
building. 

In reference to the use of forms instead of coffer-dams, the writer 
is of the opinion that this is most economical where the depths to be 
sunk are so shallow that all the concrete can be placed before sinking 
starts; then there is no delay and no increased friction. 

"Reinforced concrete construction has been used for air chambers, 
but not for small caissons. Much progress has boon made in caisson 
construction during the past few years, and is still being made. A 
firm which drops out of caisson work for three or four years will find 
old estimates of cost of very little use in bidding on new work, and 
when this is true of experienced men it is surprising to see the way 



DISCUSSION ON PNEUMATIC FOUNDATIONS 29 

that novices underbid experienced contractors. It is certainly never Mr. Tiiomson. 
to the advantage of an owner to let a contract to a company which has 
never had any experience in work of this kind, especially when the 
price bid is below actual cost. 

The writer has been asked how the "sand hogs" are paid, and in 
reply submits the following copy of the rates, terms, etc., as furnished 
by L. L. Brown, M. Am. Soc. C. E. 

An agreement made and entered into by and between the firm 

of and the International Comprossod Air Workers of 

America. Witnesseth : 

First. — That from the first day of May, 1906, 8 hours constitutes 
a day's work, on Mondays, Tuesdays, Wednesdays, Thursdays, Fridays, 
and Saturdays of each week, including 30 min. for dinner. 

Second. — That all labor performed on legal holidays, including 
Sundays, shall be entitled to an advance of 50%, whether in or out of 
the caisson. 

Third. — That the minimum rate of wages for pressuremen shall be 
as follows: Up to 50 ft., $3.50 for one 8-hour watch, including 30 min. 
for dinner. 

^3.75 for two 3-hour watches, 
3.75 for two 2-hour watches, 
4.00 for two 1 hour and ."0-min. watches, 
4.25 for two 1-hour watches, 
4.50 for two 45-min. watches, 
4.50 for two 40-min. watches. 

From starting of concreting of air chamber, 50 cents extra per day 
shall be paid. 

All depths to be measured from standard high tide, as established 
by the engineers of this port. 

Fourth. — In case a gang is called out of the caisson, each man of 
the said gang be allowed full time for the said watch. 

Fifth. — The minimum rate of wages for outside lock-tenders shall 
bo $3.50 for 8 hours. 

Sixth. — Foreman shall receive $1 per day extra, and assistant fore- 
man shall receive not less than 50 cents per day extra. 

Seventh. — That all employees shall be paid on Saturday of each 
week, up to and including the previous Thursday. 

Eighth.— The minimum rate of wages for outside work shall be 
$3.50 for 8 hours. 

Ninth. — That the firm of agrees to employ only mem- 
bers of the I. C. A. W. IT. of A., or such others as will be recognized 
by them througho^it the United States. 

Tenth. — That in case labor is required to weight down the caisson 
during the sinking, the members of the I. C A. W. U. of A. in the 
employ of the firm of shall have the preference. 

Eleventh. — If any disputes arise, notice shall be given in writing 
by the party aggrieved within 24 hours after the same. Upon the 
failure of the party notified to adjust the said disi)utes, the same shall 
bo submitted to arbitration. 

Twelfth. — All disputes shall be submitted to a joint board of arbitra- 



From 


50 to 


60 ft.. 




60 to 


70 ft.. 




70 to 


80 ft.. 




80 to 


90 ft.. 




90 to 


95 ft., 




95 to 


100 ft.. 



30 DISCUSSION ON PNEUMATIC FOUNDATIONS 

Mr. Thomson, tion. Consisting of three persons selected by the firm of 



with three members of the I. C. A. W. U. of A.; this board, failing to 
agree, shall select an umpire, whose decision shall be final and binding 
on both parties. 

Thirteenth. — That a dressing-room with hot water, soap, towels, and 
coffee (made without steam) be furnished to the men on leaving the 
caisson; the temperature of said room to be regulated according to 
the weather. Also a day and night man to take charge of said room. 

Fourteenth. — In case an employee is required to work outside, 
ample time will be given him to change his clothes after leaving the 
caisson. 

Fifteenth. — That all foremen of the I. C. A. W. U. of A. shall have 
the privilege of hiring their own men. 

Sixteenth. — That this agreement is to continue in force from the 
first day of May, 1906, until the first day of May, 1907, and if any 
change is contemplated by either party, notice in writing shall be given 
by the party desiring such change, at least three months prior to the 
expiration of this agreement. 



AMEEICAN SOCIETY OF CIVIL ENaiNEERS 

I N S T 1 T U T P: D 18 5 2 



TRANSACTIONS 



Paper No. iioo. 

THE LOW STAGE OF LAKES HURON AND 
MICHIGAN. 

By C. E, GnuNSKY, M. Am. Soc. C. E. 



With Discussion by Messrs. H. M. Chittenden, and C. E. Grunsky. 



Lakes Huron and i\richigan, if mean annual water elevation alone 
is considered, have been at less than the normal stage since 1888. 
These lakes were abnormally low in 1895 and in 1896. The Board of 
Engineers on Deep Waterways, in 1900, called attention to the de- 
pressed stage of the lake waters, giving the amount of the depression 
below what was then thought to be the normal at about 1 ft. for the 
preceding 15 years. At that time this depressed stage was attributed 
to certain natural and artificial changes that had been made at the head 
of St. Clair River, which is the outlet from the lakes. 

The importance of restoring the lakes to normal elevation and of 
holding them so high that navigation interests will be fully protected, 
covipled with the desirability of withdrawing water from the lake 
system for sanitary and inland navigation purposes in limited, yet not 
inconsiderable, amounts has prompted the following study of the effect 
of water storage in Lake Superior upon the water elevation in Lakes 
Huron and Michigan, and of the causes to which the protracted low 
stages of the past in the two lower lakes should be ascribed. 

The data herein used are taken from the published annual reports 
of the Chief of Engineers, IT. S. Army." 

* See particularly the report of E. S. Wheeler. M. Am. Soc. C. E., Assistant Engineer, 
in Annual Report of the Chief of Engineers, U. S, A., 1903, Part 4. p. 3855. 



33 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 



Fig. 1 and Plate XII show the mean annual discharge for each of 
the two rivers, the St. Mary's, which flows from Lake Superior into 
Lake Huron, and the St. Clair, which, as already stated, carries the 
outflow from Lakes Huron and Michigan, and also the water-yield 
of the drainage basins tributary to the lakes, and the mean annual 
elevations of Lake Superior and of Lakes Huron and Michigan. 

With this information, mass-curves of the water-yield of Lake 
Superior drainage basin could be constructed, and the effect of storage 
in this lake upon the discharge through its outlet, the St. Mary's 
River, could be determined. 

Some of the conclusions reached from these studies and from the 
records referred to are here briefly stated. 

TABLE 1.— Lake Stages. 
Monthly Means for Typical Years. Elevations in Feet Above Sea Level.* 
Lake Superior. 



Month. 



Jan. . 
Feb., 
Mch. 
Apr . 
May . 
June 
July 
Aug. 
Sept. 
Oct. , 
Nov., 
Dec. , 
Year, 



1861. 



602.59 
602.26 
602.12 
602.53 
603.16 
603.31 
603.47 
603.43 
603.34 
603.37 
603.03 
602.65 
602.94 



602.21 
601.97 
601.52 
602.10 
602.50 
602.52 
602. 8S 
603.. 34 
604.19 
603.67 
603.33 
602.68 
602.74 



1879. 



601.39 
600.99 
600.74 
600.87 
601.25 
601.88 
601.68 
601.71 
601.64 
601.66 
601.48 
601.08 
601.32 



1892. 



601.38 
601.02 
600.84 
600.99 
601.50 
601.86 
602.00 
602.01 
602.07 
601.96 
601.68 
601.40 
601.56 



19U1. 



602.65 
602.38 
603.13 
602.27 
602.65 
602.60 
802.97 
603.19 
603.07 
603.14 
603.08 
602.70 
608.74 



Lakes Huron and Michigan. 



Month. 


1861. 


1872. 


1876. 


1886. 


1895. 


Jan 


582.83 
582.78 
.588.92 
.582.89 
582.94 
583.18 
583.27 
583.19 
.582.99 
582.67 
5H2..55 
.581.20 
582.86 


580.99 
.580.79 
580.29 
.580.71 
581.11 
581.51 
581.61 
581.. 58 
5S1.48 
581.36 
5S1 .06 
.580.77 
.581 . 10 


581.74 
581.72 
581.85 
582.12 
.582.73 
583.82 
.583.66 
583.60 
583.49 
583.09 
.582.94 
582.75 
582.74 


.582.67 
.582.71 
.582.93 
583.22 
583.55 
583.64 
583.48 
583.33 
583.15 
.588.02 
582. t:^ 
582. 4;^ 
.583.07 


.580 03 


Feb 


579 91 


Mch 


579 92 


Apr 

May 

June 


5S(I.()2 
.5811.18 
580 26 


July 

Aug 

Sept 

Oct 

Nov 

Dec 


.580.23 
.580.14 
.580.01 
,579.74 
579. :« 
.579 Oi> 


Year 


579. <H) 







*The elevations in TabU^ 1 are not in perfect accord with the liike-stage diagram pub- 
lished by the United States Lalie Survey. They are based on elevations published in offi- 
cial reports of l!t03, 



Tirii; LOW STAGE OF LAKES HURON AND MICHIGAN 



33 



1800 








' .1 












'JJ 












i 1 1 










i. 














1 








1865 








~Tr 














[ 












J 




























[ 




isro 




1 
















1 j 


1 












1 i 


1 












.i 
















1875 








Hi, 










1 




1 i 








1 


i 


1 










1 


!| 










1 


' 1 


T 
1 






1880 










1 














1 








! 


.■ 












! . 












■J 


I ! 








1885 


















1 1 


. 


























. 




I 




















1890 


















1 














[ZLZj 














! 














1 
1 






1 


; 


1895 






. 














. 


1 
1 










1 


. 


I 




" 








.. 




OI 








!l 




! >< 


1900 








J 




p 








1 
I 


i 




^ 










-i-i 




2. 










J 


tfi 
















1905 








'1 


1 
















1 
1 




















5^ 

p CO 

W l-H 



B. P 



1800 



1865 



1870 



1873 



1880 



1885 



^ < 



1895 



p P "^ 



1900' 



1905 



34 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 



The area of the drainage basin of Lake Superior is about 76 100 
sq. miles ; the area of the lake is about 32 100 sq. miles. 

The area of the drainage basin of Lakes Huron and Michigan, 
including the Lake Superior basin, is 213 900 sq. miles, and the water- 
surface area of the two lower lakes is 45 500 sq. miles. 

TABLE 2. — Discharge of St, Mary's and St. Clair Kivers.* 
Computed from Published Monthly Means. 



Year. 



1860. . . 
1861... 
1862... 
1863... 
1864... 
1885... 
1866... 
1867... 
1868... 
1869.. . 
1870... 
1871. . 
1872... 
1873... 
1874... 
1875... 
1876... 
1877... 
1878... 
1879... 
1880... 
1881... 
1882... 
1883... 
1884... 
1885... 
1886... 
1887... 
1888... 
1889... 
1890... 
1891... 
1892... 
1893... 
1894... 
1895... 
1896... 
1897... 
1898.. . 
1899... 
1900... 
1901..., 
1902..., 
1903..., 
1904..., 
1905..., 

Means 



St. Mary's River. 



Elevation, 

in feet, 

Lake Superior. 



603 
602 
602 
602 
601 
602 
602 
602 
602 
602, 
602, 
602, 
602. 
602, 
602, 
602. 
603. 
602. 
601. 
601. 
601. 
602. 
602. 
602. 
601. 
602. 
601. 
601. 
603. 
602. 
603. 
601. 
601. 
601 
602 
602. 
602. 
602 
602 
602 
602 
602 
603. 
603 
602 
602. 



Mean discharge, 
in second-feet. 



002.31 



91 400 

92 300 

86 500 
79 900 

73 600 
81 600 

81 200 
85 800 

82 000 

88 400 

82 300 

74 600 
78 800 

83 500 
82 400 
85 300 
91 200 
82 800 
73 600 
60 600 
69 6(X) 
78 300 

77 500 
73 300 

69 900 

75 300 

70 700 
70 300 

66 300 

67 500 
65 200 
58 300 
57 400 
63 000 
73 500 

76 800 
73 HOO 

78 900 
73 300 
85 700 

87 300 
87 300 
85 000 

89 (X)0 

87 0(X) 

88 000 



St. Clair River. 



Mean elevation, in 

feet. Lakes Huron 

and Michigan. 



9(10 



582.63 
582.59 
582.56 
583.11 
581.55 
581.29 
580.95 
581.41 
580.91 
581.03 
581.92 
581.94 
580.88 
581 .35 
581.80 
581.52 
582.61 
582.40 
583.07 
581.17 
581.28 
581.79 
582.20 
583.43 
583.59 
582.79 
583.01 
582.37 
581.66 
581.21 
581.08 
580.48 
580.38 
580.62 
580.77 
579.78 
579.50 
580.12 
580.30 
580.30 
580.29 
580.55 
580.30 
580.34 
580.78 
580.91 



Mean discharge, 
in second-feet. 



581.86 



227 800 
226 800 
223 200 

213 800 
303 400 
196 900 
187 500 
198 4(X) 
187 900 

190 500 
209 500 

214 100 
193 000 
198 800 
206 600 
205 7(X) 
223 300 
218 300 
211 500 

191 500 
198 500 
205 000 
309 000 
315 800 
217 6(X) 
226 8(X) 
231 600 
217 200 
204 000 
193 500 
185 600 

176 400 
171 000 

177 700 

179 700 
164 400 
155 5(K1 
171 100 

171 200 
174 000 
173 300 

180 100 
167 300 

172 500 
182 500 
185 000 



196 400 



*The lake elevations in this fable are in substantial conformity with the diagram issu«»d 
by the U. S. Lake Survey. The discharges of St. Mary's and St." Clair Rivers are, for ihe 
most part, from the official reports already mentioned. 



PLATE XII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIIl, No. 1100. 

QRUNSKY ON 

THE LOW STAGE OF LAKES HURON AND MICHIGAN. 



LAKI^S HURON AND MICHIGAN 



583 Ft 

582" 

581 " 

580 " 
SecFt. 
240 000 

330000 

230 000 

210000 

200 000 

190 000 

130 000 

170 000 

100 000 

150 000 

1 10 000 

l:;0 000 

120 000 

110 000 

100 000 

00 000 

80 000 

roooo 

GO 000 
50 000 



1 








n 


Si 


ti 
















1 










L^ 


— 




-1 




Elevati)n >£' Vater Surface 

Mean Annual 

1903 ba um-P'lane 


















-— 


- 


- 


:;::; 








1 — ' 


- 


- 






— 


L 






:: 


= 


... 


- 


581.3f. 


- 


- 


— 


n 












... 


- 




























1 








L. 


























































-. 


























- 




"^ 


^ 


" 


-- 


SgQ.l 


L. 






1 


-_j 


- 


_J 


— 










1 
















































... 




1 
















■^ 






















1 

i 






































i 
1 












































I 

















— 












- 






1 — 




























































1 










— 1 




i 


k- 




, 




niiif.l,'nr^<. rif St, Hl.-Lir IjU-or 














! 1..'— 


i 


1 








— 1 
L_J 






Water i'ield of Entire Basin [Tributary 










1 1 
1 1 


— 










1 
1 — 







U 


1 




K-l 


r 










- 


to the Lakes incl. liak 

Same' Water Yield iii 5 

1 1 1 1 1 1 


(Siipe 
Tear 


rio'r _| — 
Periods. 


_ 


— 


1 
._L._ 








- 




I 


































I- 
1 






















r-1 






1 
1 
1 
















■" 












- 


... 




! 
























... 




... 








... 




J 


-- 


— 


— 


■ - 

1 


- 












































--, 










... 


1 


r-1 
1 






1 


_ 








._. 














._. 


















--I 




I 




""■* 


... 


















I 




















- 










— 






















1 




'^1 
1 
1 






— 


,._P_j 




._ 




















^ 


















































1 
1 













._. 




'— ' 


— 






— 1 


































— 


- 






















, 


^ 


: 

L.. 


^ 




I- 




— 








1 
1 
t 






































... 










1 
1 
1 


— 


' 


- 


i 


- 


"~i 




^ 


— 


— 








1 
1 


- 








- 


L.. 






















■■" 


■■ 


— 


^_ Tw r 


— 1 

1 
1 


4-i 




J 

1 




















1 ~ 

1 





























... 


_. 








1 L 
^Li ! 


L 


L 


L 

— 1 — 
1 










1 
















... 


L 


h-H 






— 










^-1 




,__ 


r--i— 1 












.il 


— 


... 




S|t.Clai 
Water 


rK.- 
Yitld 


of pi!B4m(ExcI.L.Sup!) 


•1-S 


on 


._, ,^ 


— 


— 




— 
.- 


— 


- 


— 




— 













._ 




— 


1 


























'' 


i 


in 








- 










- 


'- 












_ 










































































- 


- 








































































































1 




































































i 
































ri 





















34 

sq. 

inc 

sui 



I860. 
1861. 
1862, 
1863. 
1864. 
1865. 
1866. 
1867. 



1870. 
1871. 
1872. 
1873. 
1874. 
1875. 
1876. 
1877. 
1878. 
1879. 
1880. 
1881. 
1882. 
1883. 
1884. 
1885. 
1886. 
1887. 
1888. 
1889. 
1890., 
1891. 
1892., 
1893.. 
1894.. 
1895.. 
1896.. 
1897.. 
1898.. 
1899.. 
1900.. 
1901.. 
1902., 
1903., 
1904.. 
1905.. 



Mear 



byth 
mosi 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 



35 



TABLE 3.— Lake Superior. 
Elevation and Outflow Through St. Mary's Eiver. 





1860-1888. 


1889-1905. 


1860-1905. 




Ii 

9 a 


o v 


a . 

^ 

« s 


o "v 


a . 

03 tu 

3 


^ i 

o "V 
en 


Lowest monthly mean . . . - 

Highest monthly mean. . . -J 

Range of monthly means. . 
Lowest annual mean j 

Highest annual mean j 

Range of annual means 

Minimum seasonal range \ 
of monthly means, low - 
to succeeding high / 

Maximum seasonal range ( 
of monthly means, low-; 
to succeeding high ( 

Ordinary seasonal range of 
mnnthly means , , 


600.75 1 
Mar., '79 - 
Mar. '80 \ 

604.15 
Sep., '69 

3.40 

601.35 
1879 

603.00 
1876 

1.65 

0.61 

1870 

2.67 
1869 

1.28 


49 500 
Mar., '79 

116 600 
Sep., '69 

67 100 

60 600 
1879 

92 200 
1861 

31 600 


601.00 
Feb., '93 

( 603.. 50 i 
< Sep., '99 V 
( Oct .1900 \ 

2.. 50 

601.55 
1892 

602.75 
1901 

1.18 

0.68 , 
1891 

1.81 
1899 [ 

1.24 


44 700 
Mar., '92 

100 400 
Sep., '99 

55 700 

57 400 
1892 

87 200 
1901 

29 800 


[600.75 

[604.15 

3.40 
[601.. 35 

[603.00 
1.65 

0.61 
2.67 
1.26 


44 700 

116 600 

71 900 
57 400 

92 200 
34 800 






Mean, whole period 


602.26 


78 900 


602.38 


76 100 


602.31 


77 900 







TABLE 4. — Lake Superior. 

Elevation and Outflow Through St. Mary's Eiver, and Water- Yield of 
Drainage Basin for Periods of About Five Years. 



Period. 


Mean elevation, 
in feet. 


Mean discharge of 

St. Mary's River, in 

second-feet. 


Mean water-yield of 

drainage basin, in 

second-feet. 


1860-65 


602.45 
602.42 
602..38 
602.08 
602.12 
602.02 
602.04 
602.56 
602.66 


84 200 
83 900 
80 700 
75 600 
74 800 
68 000 
65 600 
79 600 
87 200 


81 400 


1866-70 


81 400 


1871-75 


87 100 


1876-80 •. 


72 300 


1881-85 


74 000 


1886-90 


68 000 


1891-95 


69:300 


1896-1900 


82 100 


1901-05 


86 900 






Means 


602.33 


77 900 


78100 







36 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 



All elevations herein noted are based on the precise levels of 1903. 

The area of the drainage basin of Lakes Huron and Michigan is 
given as 137 800 sq. miles. The water surface of the two lakes has 
an area of 45 600 sq. miles. 



TABLE 5. — Lakes Huron and Michigan. 
Elevation and Outflow Through St. Clair River, 



Lowest monthly mean -j 

Highest monthly mean 

Range of monthly means 

Lowest annual mean -J 

Highest annual mean - 

Range of annual means 

Minimum seasonal range of ( 
monthly means, low to sue--! 
ceeding" high f 

Maximum seasonal range of 
monthly means, low to suc- 
ceeding high 

Ordinai-y seasonal range of 
monthly means 

Mean, whole period 



1860-1883. 



580.05 
Mar., '69 



583.60 
Jun., '8C 



3.60 



580.88 
1873 



.583.01 
1886 



HAS 



0.34 

1879 



2.10 

1876 



1.06 



o "V 

s.a-2 



135 200 
Feb., '72 



272 400 
Jun., '86 



138 000 



187 900 
1868 



231 600 



43 700 



581.91 



208 800 



1889-1905. 






579.00 
Dec, '95 

581.80 
July, '89 

2.80 

579.50 
1896 

581.21 
1889 



0.35 I 
1895 I 



1.55 1 
1899 f 



fe 2 



1.09 



580.52 



110 600 I 
Feb. '96 f 

222 000 ( 
Aug.'89 ) 

129 500 

155 500 I 
1896 f 

192 500 I 
1888 ) 

37 000 



174100 



1860-1905. 



579.00 

583.60 

4.60 

579.50 

582.95 
3.45 

0.34 

2.10 



581 .36 



110 600 

273 400 
161 800 
155 500 

231600 
76100 



196 400 



In giving the seasonal range of monthly means in Table 5, the rise 
from a low to the following high stage alone was taken into account. 
In the season 1871-72, there was a drop in the water surface of 2.58 ft. 

The water-yield of the lake basins in the foregoing tables was 
determined as follows: In the case of Lake Superior, the annual 
water-yield is the outflow from the lake, that is, the discharge of St. 
Mary's River increased by the storage increase, or decreased by the 
storage decrease, in Lake Superior during the year. The annual 
water-yield of the entire drainage basin of Lakes Huron iiiid ]\ri(']!ifi;in 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 



37 



(including Lake Superior) is the flow of St. Clair River plus the dis- 
charge through the Chicago Drainage Canal, increased by the annual 
storage increase in these two lakes, or decreased by the storage decrease. 
The water-yield of the restricted drainage basin of Lakes Huron and 
Michigan, that is of the drainage basin exclusive of Lake Superior, is 
found by subtracting the mean annual discharge of St. Mary's River 
from that of St. Clair River and adding to the remainder the annual 
storage increase in Lakes Huron and Michigan or subtracting there- 
from the annual decrease of storage, as the case may be, and also add- 
ing the amount of water diverted into the Chicago Drainage Canal 
(which was opened in January, 1900, and takes about 4 167 sec-ft. of 
water from Lake Michigan). 

TABLE 6. — Lakes Huron and Michigan. 

Elevation and Outflow Through St. Clair River and Water- Yield of 

Drainage Basin for Periods of About Five Years. 



Period. 


Mean 

elevation, 

in feet. 


Me£in 

discliarge of 

St. Clair River. 

in 

second-feet. 


Mean 

water-yield* of 

drainage basin 

(excl. Lalce 

Superior), in 

second-feet. 


Mean 
water-yield* of 
irainage basin 

(incl. Lalie 
Superior), in 
second -feet. 


1860-65 


582.14 
581.25 
581 .49 
581 .72 
582.37 
581.87 
580.39 
580.12 
580.56 


215 300 
194 800 
203 600 
208 600 
214 800 
206 200 
173 400 
168 800 
178 300 


117 000 

118 l(jO 
124 500 
129 000 
153 300 
121 700 

95 800 
99 700 

96 500 


201 800 


1866-70 


202 000 


1870-75 


205 200 


1876-80 


204 600 


1881-85 


228 100 


1886-90 


189 700 


1891-95 

1896-1900 


161 400 
179 300 


1901-05 


183 700 






Means 


581.36 


196 400 


117 400 


195 200 







*The term, water -yield, as here used, means the total delivery of water into the 
two lakes in excess of evaporation from the lake surface. 



The normal outflow from Lake Superior through St. Mary's River, 
as determined from the records covering the 46 years, 1860 to 1905, 
is 77 900 sec-ft. The mean flow of the river during the period, 1860 
to 1888, was 78 900 sec-ft., and, from 1889 to 1905, it was 76 100 sec-ft 
The monthly mean flow of St. Mary's River ranges from about 45 000 
to about 117 000 sec-ft. 

If this lake were converted into a storage reservoir by the con- 
struction of works for the regulation of the flow of St. ^Mary's River, 



38 THE LOW STAGE OF LAKES HURON AND MICHIGAN 

it would be possible to hold the lake at or near a high stage until the 
stored water is needed to supply a deficiency in Lakes Huron and 
Michigan. The degree of benefit that can thus be secured will depend 
obviously upon the amount of water that can be stored, that is, upon 
the permissible range from low to high stage of Lake Superior and 
upon the capacity of the outlet channels. The records for the 46 
years, 1860-1905, indicate a range of mean monthly lake stages of 
3.40 ft., and a range of the annual means of 1.65 ft. The range of 
monthly means from low to high in single seasons is normally 1.26 ft., 
but this amount is frequently exceeded, and there is one season noted, 
1869, in which it reached 2.67 ft. This was an unusual fluctuation, and 
strikingly illustrates the fact that there may be a material departure 
at certain times each year from the mean annual stage. The probable 
and possible departure should be carefully studied when works for the 
complete control of the outflow from the lakes are planned. 

A description of the works for the partial control of lake stages, 
made necessary by the construction of the Lake Superior Power Canal, 
which have been in service since 1902, will be found in Transactions* 

By computation in the usual way, with recourse to the mass-curve 
of the annual water-yield of the Lake Superior basin, it can be shown 
that, under complete regulation, with a range 1.5 ft. between the 
extremes of mean annual lake elevations, there would be material 
improvement over natural outflow conditions. This amount of storage 
is equivalent to a flow of 42 600 sec-ft. for 1 year. In a succession of 
seasons such as those from 1860 to 1888, it would keep the outflow 
at a minimum annual mean of about 69 200 sec-ft. The outflow of 
60 600 sec-ft. in 1879 could have been increased by 8 600 sec-ft. 
From 1860 to 1888, there would have been no time when the mean 
annual delivery of water from Lake Superior into Lakes Huron and 
Michigan would have been less than the above indicated minimum of 
69 200 sec-ft., unless by intent to conserve water for a subsequent year. 

An examination of the mass-curve of water-yield for the entire 
period, 1860 to 1905, shows the real critical period to have been from 
1888 to 1893, in the last three years of which the discharge of St. 
Mary's River fell to 58 300, 57 400, and 63 000 sec-ft., respectively. 
A controlled storage capacity of 1.5 ft. in depth over Lake Superior 

* " The Compensating Works of the Lake Superior Power Company," by G. F. Stickney , 
M. Am. Soc. C. E., Transactions, Am. Soc. C. E., Vol. LIV, p. 346. 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 39 

would have made it possible to increase these amounts to a constant 
flow of 65 700 sec-ft. 

On the assumption that the two critical periods — the one including 
the years 1878 and 1879 in the first 20 years of the discharge record, 
and the last-mentioned period — indicate what may be expected in the 
future, it may be broadly stated that an available storage of 1.5 ft. in 
depth over Lake Superior would enable the lake outflow to be main- 
tained above a minimum annual mean approximating 65 700 sec-ft. 
Should it be found practicable to make the effective storage in Lake 
Superior 2.5 ft. between the extremes of annual mean stages, then 
the regulation of the discharge from the lake will make it possible to 
keep the minimum mean annual outflow higher than if storage be re- 
stricted to only 1.5 ft. It is found in this case, if past records be 
again examined, that the critical period in the last 48 years would have 
extended from 1876 to 1893. In this period of 16 years the regulation 
of outflow and the addition of stored water would have raised the 
minimum mean annual discharge from 57 400 to about 71 400 sec-ft. 

It is self-evident that under intelligent management the lake out- 
flow would not be kept uniform throughout any series of years nor 
even throughout any single year. The aim would be to deliver the 
stored water into Lakes Huron and Michigan in the year and at the 
season of the year when it would do the most good. 

In view of the fact that the natural channel of St. Mary's River 
will always be supplemented by power canals of large capacity and by 
the navigation canals on both sides of the river, it seems reasonable 
to anticipate that the ultimate total capacity of the lake outlets may 
reach 150 000 cu. ft. per sec. at a mean lake stage. Outflow at this 
rate is equivalent in one month to a layer of water 0.44 ft. deep on 
the surface of Lake Superior. In such a year as 1869, in which the 
lake rose 2.7 ft. in the 6 months from March to September, the lake 
elevation under complete regulation and this assumed capacity of out- 
let channels could be held down a little more than 1 ft. below the 
elevation then reached. 

On the other hand, a delivery of water in the amount named from 
Lake Superior into Lakes Huron and Michigan is equivalent to a 
layer of water, over the whole surface of these lakes, of 0.30 ft. for 
each month of such flow. Wlien it is recalled that there has been a 
year in which St. Mary's River discharged only 57 400 sec-ft., and that 



40 THE LOW STAGE OF LAKES HURON AND MICHIGAN 

there has been a mean monthly discharge as low as 45 000 sec-ft., it 
will be readily understood that an assured mean annual delivery in 
excess of 65 700 sec-ft., and a possible delivery of about 150 000 sec-ft. 
for a part of the year, will have an appreciable effect on the water 
stage of Lakes Huron and Michigan. Such regulation will not, as a 
matter of course, change the mean reduction of water level that would 
result from the continuous withdrawal, during a long period of time, 
of a given quantity of water from these lakes; but it will, under 
intelligent management, result in distributing the amount of depres- 
sion of the water surface to the several months of the year, and to 
successive years, so that the lowering of the lakes due to water with- 
drawal will be least in the months when the lakes are lowest. In other 
words, by regulating the inflow from Lake Superior, the effect of the 
withdrawal can be, at least partially, offset at low lake stages. This 
effect can thus be artificially made maximum when the lakes are 
high, when there is no injury therefrom to navigation interests, and 
can be kept at a minimum when lowering would be detrimental to 
navigation. The effect of controlled water storage in Lake Superior, 
whether it be restricted to the 1.5-ft. or the 2.5-ft. limits, will be 
unquestionably of measurable benefit. 

The desirability of utilizing Lake Superior as a storage reservoir, 
together with the utilization of the controlled flow of St. Mary's River 
to offset the effect of water diversion at Chicago for sanitary purposes, 
appears to have been pointed out first by Rudolph Hering, M. Am. Soc. 
C. E., in his report of October 15th, 1907, on the disposal of Calumet 
sewage. 

While it is apparent that in such a year as 1892 the control of 
storage in Lake Superior would have made it possible to increase the 
mean annual flow of St. Mary's River from 57 400 to about 65 700 
sec-ft., an increase of 8 300 sec-ft., or enough to raise the level of Lakes 
Huron and Michigan about 2 in. (2.4 in. less the effect of increased 
flow of St. Clair River due to greater lake elevation), it must be 
remembered that the effect of this stora^ upon the mean stages of 
Lakes Huron and Michigan for a series of years would be compara- 
tively slight. This effect, notable though it may be for a single year 
or for several years in which stored water increases the flow of St. 
Mary's River, is offset in a long series of years by the fact that, in the 
years of more than normal water production, the water delivery into 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 41 

the lower lakes must be cut down below natural flow, otherwise there 
would be no water for the improvement of conditions in those years in 
which the natural outflow from Lake Superior is small. The perma- 
nent effect of a controlled outflow from Lake Superior upon the stages 
of the lower lakes, averaged for a long series of years, would be due 
mainly to the slight modification of the outflow from these lakes result- 
ing from the modified lake elevation. This effect would be slight, and 
may prove difficult to trace. 

Whether an arrangement for storage in Lake Superior, with a 
range of annual mean elevations as great as 2.5 ft. to secure maximum 
benefit, can be effected, is not known at this time. It seems possible 
that this range might be attainable, in view of the fact that the 
extreme range of monthly means (1860-1905) has been 3.4 ft., and that, 
if the unusually high stage of 1838 be taken into account, this range 
has been about 4.5 ft. A computation of the effect of the larger 
storage upon the flow of St. Mary's River has been made, as above set 
forth, to emphasize the point th^ controlled storage in Lake Superior 
can be made beneficial to lake navigation, and, therefore, would be a 
partial offset to any water diversion from Lakes Huron and Michigan. 

Based on the water elevations for the lakes, and the amount of flow 
in the St. Mary's and the St. Clair Rivers, as noted in the reports of 
the II. S. Army Engineers, it is found that a continuous withdrawal of 
10 000 sec-ft. would lower Lakes Huron and Michigan about as 
follows : 

For a mean annual elevation of 580 ft., the lowering would 

be 0.49 ft. 
For a mean annual elevation of 581 ft., the lowering would 

be 0.47 ft. 
For a mean annual elevation of 582 ft., the lowering would 

be 0.45 ft. 
For a mean annual elevation of 583 ft., the lowering would 

be 0.43 ft. 

A continuous withdrawal of this amoxmt of water during the 
period 1860 to 1905 would have reduced the water elevations by a 
mean amount of about 0.46 ft. 

As already shown, the offset to this depression, due to storage in 
Lake Superior during a year, corresponding to a year of minimum 



42 THE LOW STAGE OF LAKES HURON AND MICHIGAN 

flow of St. Mary's Eiver, such as 1892, would be about 0.17 ft. (storage 
taken at 1.5 ft.). 

The stages of Lakes Huron and Michigan have been lower during 
the years since 1888 than during the years covered by records preceding 
that date. Not only has the mean for the whole series of years been 
lower by about 1 ft., but an extreme low mean annual stage was reached 
in 1896, which was about 1.35 ft. lower than any recorded low water 
(annual mean) during the earlier period 1860 to 1888. The year 1888 
is a convenient line of division in making this comparison, because 
the lake stage in that year was about normal, and has since remained 
less than normal. 

The mean lake elevation from 1860 to the close of 1888 was 581.89 
ft. The mean lake elevation, 1889 to 1905, inclusive, was 580.45 ft. 
The lakes were, therefore, about 1.44 ft. lower in the later 17-year 
period than in the earlier 29-year period. The mean lake elevation 
determined for the 46 years, 1860-1905, is 581.36 ft. The mean stage 
of the lakes during the 17 years, 188^-1905, was, therefore, lower than 
the mean for the entire period by 0.91 ft. 

The long-continued depressed stage of Lakes Huron and Michigan 
has been attributed by the Board of Engineers on Deep Waterways 
to the enlarged section of St. Clair River and the improvement of the 
outfall from Lake Huron into the river. On this subject, the Board 



"There is now a channel over 40 feet deep from the lake into the 
river, the increased outflow through which has lowered the general 
level of Lakes Huron and Michigan about 1 foot." 

The Board of Engineers also says:t 

"The mean level of Lake Huron is apparently about 1 foot lower 
than it was fifteen years ago, which change has resulted from the 
enlargement and deepening of channels for waterway improvements 
and from the natural erosion of the bed of the river at the outlet of 
the lake." 

The general depth of the foot of Lake Huron, li miles above the 
head of St. Clair River, is stated by the Board of Engineers to have 
been originally from 21 to 27 ft., with numerous shoals 16 to 18 ft. 
deep. A channel, 2 400 ft. wide and 21 ft. deep, at a mean stage, has 

* " Report of the BoariJ of Engineers on Deep Waterways between the Gi-eat Lakes and 
the Atlantic Tide Waters," 1900, p. 37. 

t " Report of the Board of Engineers on Deep Waterways between the Great Lakes and 
the Atlantic Tide Waters," l'.»00, p. H3 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 43 

been cut through these shoals. Concerning the deepening of water 
in the head of St. Clair River by the scouring action of the water, the 
Board of Engineers states that, in 1867, surveys showed the depth of 
water on the bar over which the lake discharges into the river to have 
been 27 ft., and the central depth of water in the gorge at the head 
of the river to have been 48 ft. The surveys of 1898 and 1899, according 
to the Board's report, showed that a channel had been scoured through 
the bar to a depth of 75 ft., and that the water depth in the gorge at 
its narrowest point was from 48 to 66 ft. 

The 15 years, 1885 to 1899, inclusive, to which the Board of Engi- 
neers referred in making its comparison of lake elevation in recent 
years with former elevations, show a mean water-surface elevation 
which was 0.95 ft. lower than that of the 25 years, 1860 to 1884, 
inclusive. 

The possibility that altered conditions at and near the head of the 
St. Clair River had the effect of modifying, to some extent, the stage 
of Lakes Huron and Michigan must be admitted, but changes of out- 
let capacity have probably been only minor factors in producing the 
low lake stages, if indeed they have been of any effect. This will 
appear from the following consideration: 

It must be apparent that the lowering resulted either from the 
cause to which it is attributed by the Board of Engineers, or there 
must have been less water presented by Lake Huron for delivery 
through the St. Clair River. Both causes may have contributed to the 
result. If the depressed water surface can be accounted for in part 
by a decreased water production of the drainage basin which is 
tributary to St. Clair River, then only the remainder, if there be any, 
will be ascribed to the increased outflow capacity of the head of St. 
Clair River. 

That there has been a deficient water production in the drainage 
basin treated as a unit of the three lakes, Superior, Huron, and 
Michigan, during a long period subsequent to 1888, is primarily indi- 
cated by a decrease of the St. Clair River discharge. This decrease 
is quite as noticeable as the depressed lake elevation, as will appear by 
inspection of the diagrams on Plate XII. The flow of the St. Clair 
River, however, does not by itself represent the water production of 
the lake drainage basin from year to year, because some of the water 
remains stored in the lakes. It will be instructive, therefore, to 



44 THE LOW STAGE OF LAKES HURON AND MICHIGAN 

determine the annual water quantities in excess of evaporation received 
by the two lakes, Huron and Michigan, from all sources, and to com- 
pare these quantities with each other for the same time periods for 
which comparisons of lake elevations have been made. 

The net quantity of water thus annually received by the two lakes 
is found by adding to the annual outflow through the St. Clair River 
the storage accretion (plus or minus as the case may be) of the two 
lakes, and adding also the diversion through the Chicago Drainage 
Canal, which has been about 4 200 sec-ft. since January, 1900. 

The result of this computation, based on the figures contained in 
the Wheeler and other U. S. Engineer reports, shows conclusively that 
there has been a very decided falling off in the water-yield of the 
Huron-Michigan drainage basin in recent years, and that the drop in 
the water-surface elevation of the lakes is coincident with this decrease 
of yield. The water-yield of the basin (run-off, and rain on the lakes, 
less evaporation), _ as computed, noted for periods of about five years 
and expressed in second-feet continuous flow, has already been noted 
in Table 6 and is shown on Plate XII. 

For the 17 years, 1889 to 1905, the entire drainage basin of Lakes 
Huron and Michigan, including Lake Superior, contributed to the 
lakes a mean flow of 175 000 sec-ft. of water, whereas the normal water 
yield of the basin is about 195 200 sec-ft. The deficiency of 20 200 
sec-ft. is enough, at the normal stage of the lakes, to account for about 
0.94 ft. of deficient water elevation. This being the case, the lakes 
would have been, at a mean elevation, nearly 1 ft. lower than normal 
subsequent to 1888, even though outlet conditions from Lake Huron 
had not changed. 

It has already been shown that the mean elevation of the two lakes 
for the period 1860-1905 was 581.36 ft. and for the period 1889-1905, 
580.45 ft. Had there been no withdrawal of water from the lakes at 
Chicago, these elevations would have been 581.38 and 580.50, re- 
spectively. The lakes would have been 0.88 ft. lower, in the 17-year 
period following 1888, than their normal elevation. This depression 
of 0.88 ft. is the combined effect of less than normal rainfall (in part 
perhaps more than normal evaporation) and of changes at the head 
of the St. Clair River. As above set forth, however, the effect on 
lake elevation that might be reasonably attributed to the first cause 
alone, viz., climatic conditions, is about 0.94 ft., slightly in excess of 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 45 

the actual depression, leaving no drawing-down effect to be ascribed 
to the changes at the head of St. Clair River. 

The conclusion that the channel changes at the head of the St. 
Clair River have not materially affected the stage of the lakes, if at 
all, is, as already stated, based on the lake elevation and discharge 
figures published by the United States Army Engineers. It is note- 
worthy that these figures show that the material reduction, after 1888, 
in the water productiveness of the drainage basin of Lakes Huron and 
Michigan, was confined almost entirely to the area directly tributary 
to these two lakes. There was no pronounced falling off noted in the 
discharge reaching these lakes through St. Mary's River from Lake 
Superior. However, as the published stream measurements and river 
stages for the same period of 17 years show a decrease in the discharge 
of St. Lawrence River, the outlet of Lake Ontario, amounting 
(after correction for depletion of lake storage and the diversion at 
Chicago) to about 18 000 sec-f t. — nearly as much as the decreased 
water production of the area tributary to the St. Clair River, 20 200 
sec-ft. — there would appear little room for doubting the substantial 
accuracy of the published discharge tables of the St. Clair River, and, 
therefore, the conclusion, as stated, relating to the main cause of the 
low lake stages, seems to be based on reliable premises. 

That this conclusion, which attributes the low lake stages of Lakes 
Huron and Michigan to a long period of deficient water production 
in the Huron-Michigan basin, and in a very small degree, if at all, 
to changes at the head of the outlet channel, is probably correct, is 
borne out by a statement of E. E. Haskell, M. Am. Soc. C. E., in his 
report of July 16th, 1900,* to the effect that what had a year previously 
been reported as a clear case of enlargement by scour of the head of 
St. Clair River was based on a comparison of a preliminary survey 
of 1898, with a chart based on a survey in 1867, and showed an 18-ft. 
cut over a portion of the gorged reach. Mr. Haskell goes on to say 
that an older chart of 1859 was subsequently found to agree much 
better with the survey of 1898. He caused old notes to be re- 
platted, and careful comparisons based thereon have led him to the 
conclusion that the changes have been small. Between 1859 and 1867, 
the most restricted section may have been enlarged 9 000 sq. ft. 



* Report of Chief of Engineers, U. S. A., 1900, PI,. 8, p. 5323. 



46 THE LOW STAGE OF LAKES HURON AND MICHIGAN 

Between 1867 and 1900 the changes, if any, Mr. Haskell says, are 
unimportant. 

On the assumption that the discharge of St. Clair River has been 
approximated with a fair degree of accuracy for each year of the 
entire period, 1860 to 1905, it appears reasonably certain that the low 
stage of Lakes Huron and Michigan will not persist, that normal 
weather conditions will restore about 1 ft. of the lost mean lake eleva- 
tion, and that the only depression below the original normal will be 
a small amount, if any, due to the St. Clair River changes and a 
small amount due to the Chicago diversion. Until the Chicago 
diversion is increased above the present amount of about 4 200 sec-ft., 
a mean lake stage at about 581.20 ft. is to be expected. This, it may 
be stated, is l.YO ft. higher than the lowest mean annual lake elevation 
(579.50 ft.) of the past 48 years, which occurred in 1896. 

This conclusion relating to a probable future higher lake level 
than that of the period subsequent to 1888 is inevitable, because it 
may be accepted as a certainty that the unusual climatic conditions 
which, since 1888, have resulted in a deficient output of water from 
the Huron-Michigan basin will not continue indefinitely. The occur- 
rence of another period, as protracted as the one subsequent to 1888, of 
small water-yield in the lake basin is highly improbable. It is proper, 
therefore, to assume that the years in which such low levels as those 
of 1895 and 1896 will occur will be few and far between. 

What has occurred in the past, however, may occur again; there- 
fore means should be sought to keep up the elevation of Lakes Huron 
and Michigan during any future periods of deficient precipitation in 
the lake basin. 

To some extent the raising of the level of these lakes can be 
accomplished with controlling works in the head of Niagara River. 
Such works have been recommended, and designs therefor have been 
made by the Board of Engineers on Deep Waterways. This Board 
proposes a submerged weir 2 900 ft. long and a series of sluice-gates, 
13 in number, and each 80 ft. wide. These works, it is stated, would 
raise the level of Lake Erie 3 ft., the level of Lake St. Clair about 
2 ft., and the level of Lakes Huron and Michigan about 1 ft. No 
attempt has been made to check this forecast. It is possible, more- 
over, to go a step farther and to provide works for the throttling of the 
waterway, or for the complete control of flow, within fixed limits, of 



THE LOW STAGE OF LAKES HURON AND MICHIGAN 47 

the St. Clair Kiver. The stage of Lakes Huron and Michigan could 
then be brought under adequate control, and the lake stage could be 
maintained at all times as high as required by navigation interests. 

That regulating works will sooner or later be constructed, in the 
rivers draining the several lakes, of such a character that the elevation 
of the lakes under proper management cannot fall below predetermined 
minima, is reasonably certain. It follows, therefore, in view of the 
benefit that will be derived from the use of Lake Superior as a storage 
basin, and of the control of water surface that will result from works 
below Lake Huron, that navigation interests will be but temporarily, 
if at all, affected by a diversion of water at Chicago, or elsewhere 
within reasonable limits as to amount, and that any deleterious effect 
will not continue beyond the time when the water elevation of this 
lake and Lake Michigan will be controlled by works below Lake 
Huron. 

The stages of Lake Ontario and of the St. Lawrence Eiver are ques- 
tions apart from the one that has here been discussed. Enough has 
been said, however, to show that here, too, the less than normal stages 
of recent years are to be ascribed to climatic conditions, supplemented 
in a very slight degree only by the diversion at Chicago. The ultimate 
effect of the diversion of water from any of the lakes, offset by the 
equalizing effect of lake regulation, upon the stage of water in the 
St. Lawrence Kiver is a study that will have to be made when the 
regulating works come under consideration. 



48 DISCUSSION : the low stage of lakes HURON AND MICHIGAN 

DISCUSSION. 



Mr.chitten- H. M. Chittenden, M. Am. Soc. C. E. (by letter). — As a result of 
his investigation of reservoir possibilities in the arid regions in 1897, 
the writer became deeply interested in the subject of reservoirs in 
general as regulators of stream flow, and particularly in the great 
natural system of the St. Lawrence Basin. With the assistance of 
James A. Seddon, M. Am. Soc. C. E., a mathematician of exceptional 
ability, he undertook a study of the general problem of the interrela- 
tion of reservoirs in a descending series like those of the Great Lakes, 
where each unit, except the upper one, receives an independent supply 
from its own water-shed and a transmitted supply from the reservoirs 
above.* 

This study, the first of its kind ever made, led to the formulation of 
certain principles controlling the action of such reservoirs upon their 
outlets and upon each other, and also to the enunciation of certain 
conditions and limitations which must govern in any scheme for the 
control of the reservoir levels. The Board of Engineers appointed 
under the Act of Congress of June 4th, 1897, to investigate and 
report on a deep-water route from the Lakes to the Seaboard was then 
just beginning its work, and the late George Y. Wisner, M. Am. Soc. 
C. E., a member of the Board, criticized quite severely some of the 
writer's conclusions, but apparently coincided with them later, in the 
* main, as shown in his own report to the Board.f 

At the time of the writer's studies the question of lake level regula- 
tion was very prominently before the public. The low stage to which 
some of the lakes had fallen, due to a series of unusually dry years, 
had occasioned a good deal of anxiety among commercial interests on 
the lakes, and many propositions were put forth as to possible cor- 
rectives of the abnormal conditions then existing, and even for the 
improvement of normal conditions. It was claimed that the fluctuation 
of the lake levels could be limited to a range of 6 in. or less, that the 
levels of all the lakes could be permanently raised, and that Lake 
Superior could be utilized as a storage reservoir to help maintain 
the levels of the lakes below, and particularly to compensate for the 
prospective loss through the Chicago Drainage Canal then nearing 
completion. A scheme for the control of Lake Erie, frequently pro- 
posed, was a long-crested weir around the head of Niagara River, the 
development of the crest being sufficient to regulate the discharge 
automatically so that the surface of the lake would never rise or fall 
beyond certain prescribed limits. It was also held by some that the 
regulation of Lake Erie would react through the channels above it and 
regulate the levels of Lake Michigan-Huron in like manner. 

* Transactions, Am. Soc. C. E., Vol. XL, p. 855. 
+ Report of the Board, p. 274. 



DISCUSSION: THE LOW STAGE OF LAKES HURON AND MICHIGAN 49 

Among the conclusions arrived at in the writer's paper were the Mr. Chitten- 
following: *'^"- 

That the regulation of the fluctuation of the levels of the upper 
lakes to a limit of 6 in. was a physical impossibility on account of the 
great volume of water lost during the summer through evaporation; 

That any material restriction of the annual fluctuations of the lake 
surface tvas of doubtful practicability, because such restriction could 
only be had at the expense of uniformity of flow in the outlets, and 
this would probably be inadmissible from considerations of navigation ; 

That the periodic, or cyclic, fluctuation of level, extending over a 
series of years, could be eliminated altogether and the lakes could be 
kept from falling below a definite level that might be established; 

That the controlling works in the Niagara could not be in the form 
of a fixed weir, automatic in its action, but must be in the form of a 
movable weir by which the flow through it would be under human con- 
trol, and adjustable to the daily conditions of wind and supply of water 
to the lakes; 

That control of Lake Erie would not suffice for the control of Lake 
Michigan-Huron ; 

That the compensating effect of storage in Lake Superior for any 
permanent diversion through the Chicago Drainage Canal, or for a 
permanent enlargement of the outlets, was visionary and impracticable. 

As above stated, the first three of the foregoing conclusions were 
practically recognized in the report of the Deep Waterways Board. 
As to the fourth conclusion, the Board found that raising the level 
of Lake Erie 3 ft. would permanently raise Lake Michigan-Huron 
about 1 ft. In regard to the use of Lake Superior as a storage reser- 
voir, the Board was silent, and was therefore presumably of the opinion 
that such use to compensate for permanent diversions from the lakes 
below was impracticable. The present International Waterways Com- 
mission recently made a special investigation of the effects of the 
Chicago Drainage Canal diversion upon the levels of the lakes and of 
the means of compensating for the diversion. Nothing is said in their 
report as to storage in Lake Superior, and they likewise apparently 
considered the scheme impracticable. Since the date of the above- 
mentioned reports, Rudolph Hering, M. Am. Soc. C. E., has come out in 
strong advocacy of this scheme,* and now Mr. Grunsky's very able 
paper is under consideration. 

That two such competent authorities should advocate the plan has 
led the writer to review the reasoning by which his own conclusion 
was arrived at. He is still unable to see how storage in Lake Superior 
can compensate for a permanent diversion at Chicago or at any otlier 
point, or for a lowering due to an increase in the dimensions of any 

* " Report on the Disposal of Sewage from Calumet Subdivision of the Sanitary District 
of Chicago,'" October 15th, 1907. pp. 28-31. 



50 DISCUSSION : THE LOW STAGE OF LAKES HURON AND MICHIGAN 

Mr. Chitten- of the Outlets. The reason for this is that the mean levels of the 
lakes are dependent upon two things: the character of the outlets and 
the supply of water, using this last term in its algebraic sense to in- 
clude the negative effects of evaporation and diversions. The outlets 
remaining the same, any permanent diminution of supply must result 
in a permanent lowering of mean level, and the only way to prevent 
this, without modifying the outlets, would be to compensate for this 
loss of supply. Can storage in Lake Superior accomplish this purpose? 
Manifestly not, because the total supply upon which the mean level 
depends cannot be affected by any manipulation which can be made 
of a portion of that supply, so long as its quantity remains unchanged. 
Storage in Lake Superior cannot affect the total supply to Lake 
Michigan-Huron in the least. Any increase in storage above the 
normal can be accomplished only by restricting the outlet discharge 
below the normal, and this must result in a proportionate lowering of 
the levels of the lakes below while the accumulation is going on. When 
the storage is run out, it can do no more than make up for the loss 
which its previous withdrawal had occasioned. Of course, if the 
storage is accumulated very slowly through a long period of time and 
then run out very rapidly through a short period, a greater increase 
of gauge height will result than the previous decrease; but, if the 
durations of the diminished and increased stages are considered, the 
account will balance. Whatever may be done with Lake Superior stor- 
age, therefore, the supply to Michigan-Huron cannot be increased; 
and the diversion at Chicago thus represents a permanent and uncom- 
pensated loss. 

The storage in Lake Superior can be manipiilated so as to reduce 
fluctuations of level in Lake Michigan-Huron, but only at the cost of a 
greater increase in the fluctuations of the upper lake. How far the 
advantage in one case may be offset by the disadvantage in the other 
should control in determining the extent to which such regulation 
should be applied; but the redviction of annual fluctuations of level 
in Lake Michigan-Huron is a very different thing from compensating 
for a permanent loss of supply so as to maintain the mean level of 
the lakes. Manipulating the storage of Lake Superior cannot affect 
in the least the supply to the lakes, and is therefore powerless to affect 
their mean level or compensate for diversions. In fact, the writer 
does not understand that Mr. Grunsky really claims this, though his 
language leaves some doubt in the writer's mind as to his exact 
meaning. 

Clearly, a permanent diversion from the lakes, at any point, of a 
given quantity of water or an enlargement of the navigable channels 
by artificial means can be compensated for only by restricting the flow 
through the outlets below. Theoretically, this is perfectly feasible; 
practically, it is diflBcult of accomplishment owing to the conditions 



DISCUSSION : THE LOW STAGE OF LAKES HURON AND MICHIGAN 51 

imposed by navigation. In Lake Superior the difficulty is much less Mr. Chitten- 
because commerce has to pass through a lock anyway, and controlling ''*^°' 
works in the river would not interfere with navigation; but in all the 
outlets below the case is different. Controlling works which should 
diminish the normal discharge would tend to concentrate the slope at' 
the site of the works, increasing the velocity at such points, while the 
diminished flow might leave deficient depths in the channels below. 
These drawbacks to navigation might prove too great to be readily 
overcome. To construct works of such character and magnitude as to 
maintain the necessary slopes and depths in these channels during 
periods of restricted flow would assuredly be very costly. 

There is, however, a considerable period of each year when naviga- 
tion is suspended, and there is no obvious reason why, during this 
period, the flow of the outlets may not be cut down to whatever point 
would be necessary to raise the levels of the lakes to the desired 
heights before the next navigation season opens; but if the regulated 
mean level were elevated above the normal mean level, the increased out- 
flow during the navigation season, in the absence of works regulating the 
slope and permitting a partial closure of the channels at such times, 
would probably increase the subsidence of levels during such periods 
as compared with that under normal conditions. 

That the mean levels of all the lakes can be permanently raised; 
that the periodic or cyclic fluctuations can be eliminated; that some 
reduction can be made in the annual fluctuations ; that a much greater 
diversion at Chicago than 10 000 cu. ft. per sec. can be compensated 
for, by restricting the flow through the outlets, are measures which the 
writer believes to be within the resources of river engineering. As in 
any other project, the real question is one of cost in relation to the 
benefits to be received. 

C. E. Grunsky, M, Am. Soc. C. E. (by letter). — It is stated in the Mr. Grunsky. 
paper that the conclusion reached with reference to the small effect 
upon the stages of Lakes Huron and Michigan, that has hitherto re- 
sulted from channel enlargement at the outlet of Lake Huron, is based 
on the assumption that the discharge from these lakes, as far back 
as 1860, has been determined with a fair degree of accuracy by the 
United States Engineers. 

The rating tables, based on measurements made in recent years, and 
the observed lake stages, are the basis of the discharge tables. If any 
material channel enlargement, natural or artificial, preceded the gaug- 
ings on which the rating tables are based, then the application of these 
tables to water stages under original conditions would give too large 
results. Can the higher stages of Lakes Huron and Michigan preceding 
1888 have been possible with less outflow than indicated by the record ? 
This question has already been answered by reference to the fact that 
the interpretation of the stages of the St. Lawrence Eiver shows 



53 DISCUSSION : the low stage of lakes huron and Michigan 

Mr. Grunsky. Substantially the same decrease of flow in recent years as the St. Clair 
Eiver. It remains to be added that, if too great a flow has been 
recorded for the earlier period, the error would be a progressively 
decreasing one from 1860 to the time of the gauging. The increasing 
water yield for successive 5-year periods from 1860 to 1885, as shown 
in the last two columns of Table 6, negatives the probability of any 
such error. 

The statement is clearly made in the paper that the control of the 
outflow from Lake Superior would not change the mean reduction of 
lake level that would result from the withdrawal, during a long period 
of time, of water from the lakes. In other words, the controlled outflow 
from Lake Superior will not appreciably modify the normal stage 
of the lower lakes; but it is claimed that such control would be bene- 
ficial to navigation interests, because thereby the lakes would be pre- 
vented from dropping as low as they would otherwise go. 

Colonel Chittenden has clearly pointed out some of the difficulties 
that must be overcome in putting into practical effect a regulation of 
the outflow from the lower lakes; but the regulation can be effected. 
It is, at any rate, as pointed out in the paper and as restated by Colonel 
Chittenden, theoretically feasible to modify the lake stages beneficially 
by controlling the outflow, and to this fact attention must be directed 
when the question of setting a limit upon the amount of water that 
may be diverted from the lakes is under consideration. 



AMERICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 185 2 



TRANSACTIONS 



Paper No. iioi 

THE FLOODS OF THE MISSISSIPPI DELTA: 

THEIR CAUSES, 

AND SUGGESTIONS AS TO THEIR CONTROL/ 

By William D. Pickett, M. Am. Soc. C. E.* 



One of the great engineering problems of the age, if not the 
greatest, is to obtain such control of the waters of the Mississippi 
River as to prevent those periodical overflows which, in the past, have 
devastated and made desolate the entire area of its delta. This delta 
is probably the most extensive in the world, and has been formed, 
during past ages, by the discharge of the waters of the greatest river 
in the world, certainly the longest, in miles, and with a volume of 
discharge not exceeded by any, unless it be the Amazon, in South 
America. 

From a geological standpoint, the most plausible theory in regard 
to the formation of this delta is that at one time the Gulf of Mexico 
extended as far north as the confluence of the Mississippi and Ohio 
Rivers; that, in past ages, by the discharge of the silt brought down 
in the annual floods, the land has been gradually formed and extended 
out into this inland sea, until, about the beginning of the nineteenth 
century, it had assumed the form and area existing to-day. 

When it is considered that the level of the low-water stage at 
Cairo is about 250 ft. higher than the mean tide level of the Gulf, the 

*As a matter of interest, the fact is here mentioned that Mr. Pickett has been a Member 
of this Society since July 6th, 1853, having joined it during the first year of its existence. 
—Secretary. 



54 THE FLOODS OF THE MISSISSIPPI DELTA 

centuries of time required to develop this delta and the depth of its 
soil of almost illimitable fertility may be realized. 

Following a natural law, the land on the banks of this river is 
higher than that farther inland. The same conditions obtain on Deer 
Creek and Sunflower River, in the Yazoo Basin, as on the Bayou 
Plaquemine and other large bayous draining the lower portion of 
this delta. A cross-section of the Yazoo Delta, near the latitude of 
Greenville, Miss., made by the Engineer of the Yazoo Levee System 
previous to 1861, developed the fact that the level of the bank of the 
main river was about 13 ft. higher than comparative levels 20 or 25 
miles inland near Deer Creek and Sunflower River; such was the 
writer's information at the time. 

As a natural sequence to these conditions, the "high-lying lands," 
on the main stream and such subsidiary streams as those noted above, 
were first brought under cultivation, and, being of unexcelled fertility, 
produced the heaviest yield of cotton, corn, and other crops suitable 
to the climate. 

The area of these "high-lying," and, of necessity, more valuable 
lands, is a small percentage of the area of this delta. A much larger 
area consists of "low-lying lands," designated by the TJ. S. Geological 
Survey as "Sharkey clay," which are more or less subject to annual 
overflows, and are covered with a heavy growth of timber, such as 
white oak, red oak, water oak, cypress, and gum, all more or less 
valuable for commercial purposes. 

The cost of clearing off this heavy growth of timber and preparing 
the ground, by ditching, etc., for cultivation, together with the almost 
certainty of periodical overflows, has heretofore and will hereafter 
prevent these otherwise valuable lands from being brought under 
cultivation and rendered productive until the levees on the main river 
have been perfected so as to render these overflows improbable, if not 
impossible. 

These low-lying "overflow lands" are represented, by officers of the 
Geological Survey and by intelligent planters, to be of as great fertility 
and as productive of corn and cotton as the "high-lying" lands on the 
streams, where the ground has been properly prepared. With land of 
this character, protected from overflow, it is believed that the value of 
its timber for commercial purposes will go far toward meeting the 
cost of preparing the <fn)un(l for crops, if it docs not equal or surpass it. 



THE FLOODS OF THE MISSISSIPPI DELTA 55 

These facts being substantially true, it is proposed to investigate 
and ascertain approximately the value, to the nation, of these "over- 
flow lands" on the supposition that they could be freed from that 
incubus of "overflow" so much dreaded by the Mississippi River 
planter. 

The most reliable information as to the extent of this entire delta 
is obtained from the Agricultural Department, through Mr. J. A. 
Bonsteel, of the Bureau of Soils, who, from the most reliable records 
accessible, estimates this area at 30 526 sq. miles. From the surveys 
of that Bureau in the Yazoo Delta, it is ascertained that 65% of the 
total area is "overflow" land, leaving 35% for land under cultivation. 
Adopting this basis for the entire delta, it indicates the total area to be, 
in round numbers, 30 000 sq. miles, and the area of "overflow" lands 
to be 20 000 sq. miles of land of the highest fertility, yet it is of no 
value, and never will be, for crops, for causes before described. • 

From the writer's knowledge of the conditions in this delta, it is 
believed that the land under cultivation is nearer 25% of the total 
area than 35 per cent. The 35%, however, has been adopted in the 
following estimate furnished by Mr. Bonsteel : 

"Area of Mississippi Delta, in Square Miles, by States: 

"Total "Area of 

"State. area. overflow land. 

"Kentucky 150 100 

"Tennessee 398 260 

"Missouri 3111 2 022 

"Arkansas 5 406 3 513 

"Mississippi 6 966 4 425 

"Louisiana 14 495 9 425 



"30 526 19 745." 

In making an estimate of the yield of this now non-productive land, 
the area of Kentucky, Missouri, and 100 sq. miles of Tennessee is 
assigned to corn; 160 sq. miles of Tennessee and all of Arkansas and 
Mississippi are assigned to cotton, and all of Louisiana is assigned to 
sugar and rice. As the necessary statistics for the yield of Louisiana 
lands have not been attainable, a yield of $35 per acre has been 
estimated, the same as the yield of col^on lands. 

To compensate for lands not available for crops in the "over- 



56 THE FLOODS OF THE MISSISSIPPI DELTA 

flow" area, such as that occupied by rivers, bayous, and creeks, 10% 
has been deducted. The results, in round numbers, are as follows : 

Corn land 2 000 sq. miles = 1 280 000 acres. 

Cotton land 7 300" " =4 672 000 " 

Sugar and rice lands. . . 8 500 " " =5 440 000 " 

In placing a value on crops from these lands, the yield of corn 
has been taken at 35 bushels per acre, and the value at 50 cents 
per bushel. The yield of cotton has been taken at i bale per acre and 
the value at $40 per bale. The yield of sugar lands has been taken at 
$35 per acre, the same as cotton lands. 

Turning these statistics into dollars and cents, we have an estimate 
— conservative, it is thought — of the value of crops that can be 
realized from land that never has had and never will have much 
intrinsic value for agriculture until freed from that overhanging 
incubus — the periodical overflows from that mighty river. The little 
benefit that has been realized in the past, from this overflow land as 
a range for stock, has been much more than counterbalanced by the 
destruction of stock in such overflows. 

1 280 000 acres of corn land at 35 bushels per acre = 

44 800 000 bushels at 50 cents $22 400 000 

4 672 000 acres of cotton land at i bale per acre = 

4 088 000 bales at $40 per bale 163 520 000 

5 440 000 acres of sugar land at $35 per acre 190 400 000 

Total value of crops $376 320 000 

Bear in mind that this estimate is based on the yield of fresh land, 
which should be much greater than that from the old plantations of 
this delta which have been under continuous cultivation for 50 or 60 
years, or more. Bear in mind, also, that it is based on what the ground 
will bring forth with average cultivation, not on the profit to the 
planter. Bear in mind, also, that this estimate is exclusive of the 35% 
(9 000 sq. miles) of this delta which is supposed to be under cultivation 
and already producing crops. In years of widespread overflows on this 
delta, the crops on this area of 9 000 sq. miles will be a total loss to 
the nation, increasing the above'estimate of $376 320 000 by the loss of 
the crops on that 5 700 000 acres of cultivated land. 



THE FLOODS OF THE MISSISSIPPI DELTA 57 

The annual loss to the United States appears to be $370 000 000 
in a small area of its territory, for want of protection against the 
overflow of this river. 

As an indication that tlie writer's views of the value to the nation 
of this small fraction of its area are within conservative lines, an 
extract is given from the conclusion of an address before the Com- 
mercial Club of Cincinnati, in December, 1903, by B. M. Harrod, 
Past-President, Am. Soc. C. E., one of the oldest and ablest members 
of the Mississippi River Commission, bearing directly on this subject : 

"Now, this is the proposition : There is an area of 20 000 000 acres 
of the most fertile land in the world. At least three-quarters of it are 
susceptible of the highest cultivation. Its potential products are 
diversified; including wheat, corn, cotton, sugar and rice. The timber 
wealth of the valley is immense, and it is in every way favorable from 
climate, soil and means of transportation for development by the 
farmer, the manufacturer, and for railroads. Without levees it is a 
jungle; an uninhabitable swamp. With levees each and every acre 
can be protected at a cost not exceeding three dollars, and this protec- 
tion can be maintained at an annual cost of ten cents per acre. Nearly 
two-thirds of the work is now done." 

The prospective, addition of 4 000 000 bales of cotton and a pro- 
portionate yield of sugar — both among the necessities of life — to the 
resources of the country, renders the subject one of National im- 
portance. It seems to be a matter of paramount importance, and 
demands an investigation as to the causes of this condition of things, 
and to ascertain if means cannot be suggested that will, at least, 
mitigate the effects of these periodical disasters. 

The magnitude of this problem should suggest diffidence in ap- 
proaching its solution. In the past, it has engaged some of the best 
and most experienced engineers of the country; whereas they have 
worked along proper and conservative lines, the means at their com- 
mand have been entirely inadequate for carrying out their plans. At 
the beginning of 1861, the levee systems below Memphis were in good 
condition, as far as constructed. The ravages of war, however, and the 
unrestrained power of the river, exerted in some very high overflows, 
swept off almost the last vestige of the levees existing at that time. 

Since the close of the conflict (1865) the States bordering on this 
river have made remarkable progress in the rebuilding of the levee 
system, until at the present time, there is substantially a continuous 



58 THE FLOODS OF THE MISSISSIPPI DELTA 

system of levees on each bank from Cape Girardeau to a point below 
New Orleans. Heretofore, each State has controlled its own levee 
system, but with means entirely inadequate to the magnitude of the 
work. It is believed there should be one controlling head for the 
entire levee system of the delta, and there should be ample means to 
carry out its plans. It is believed, moreover, that in a few years 
Congress will recognize the national importance of this work, in con- 
nection with the plans for the deep-water navigation of this great 
inland sea, and provide ample means for carrying out the plans 
determined on by this central control. 

The writer has had especial opportunities for the study of the 
conditions at both ends, as he conceives it, of this great problem, and 
desires to record, in a tentative way, the views and opinions that have 
impressed themselves upon him. From 1856 to 1861, as a civil engi- 
neer, he had charge of the construction of the Memphis and Ohio, one 
of the principal railroads leading from Memphis, Tenn. Subse- 
quently, 1867 to 1873, after the ravages incident to the war, he had 
charge of the reconstruction of the same road. During that period 
he witnessed a great many of the floods which devastated the delta. 
In 1866 he was a cotton planter in the Yazoo Basin, and was a 
sufferer from the flood of that year. During all this time, with the 
predilections of a civil engineer, he naturally studied all questions 
bearing on the causes of these floods. He sought information from all 
accessible sources; from old and experienced river men — the captains 
and pilots of steamboats — from cotton planters, and from civil 
engineers. 

During these periods of overflow, this delta became an inland sea, 
40 miles wide opposite Memphis and from 70 to 80 miles wide opposite 
the Yazoo Basin, with an occasional island where spots of dry land 
on the plantations were in evidence. In the river channel, its waters 
were extremely muddy — angry-looking, awe-inspiring and repulsive. 
This indicates the extent of the writer's opportunities for forming 
correct opinions at the lower end of this problem. 

It happened that in 1876 circumstances drifted him to the North- 
west, where 28 years of his life were spent in Montana and Wyoming. 
Seven or eight months of each year for the first eight years were 
spent in the mountains at the heads of the various tributaries of the 
Missouri and Columbia Rivers, whence come the melted snows that 



THE FLOODS OF THE MISSISSIPPI DELTA 59 

make up principally the "June rise" that occasionally has such a 
potent influence in the devastation of the Great Delta. 

Twenty years were afterward spent on a cattle ranch on the 
Upper Grey Bull Eiver, at an elevation of nearly 7 000 ft. above tide. 
Within ten miles to the south is Franc's Peak, the elevation of which 
is given by the Geological Survey as 13 300 ft. above sea. On each 
side of the valley are mountain ranges with elevations of 10 000 and 
11 000 ft. 

During those twenty years' residence in tliis locality, a daily record 
was kept of temperature and precipitation, especially of snowfall. The 
necessity for irrigation called attention to all conditions influencing 
the water supply. The minimum temperature during the winter 
months varied from 24° to 47^° below zero. The annual snowfall 
varied from 50 to 110 in., the latter being unusual and extreme. The' 
average snowfall per year was from 75 to 80 in., the average annual 
precipitation being about 13 in. 

The conditions obtaining in this locality, as regards temperature, 
snowfall, and snow melting, may be safely taken as those on all 
the tributaries of the Missouri River of the same altitude. This 
altitude is about the upper limit of even forage crops. Basing opinions 
and views on the observations and knowledge gained by a long resi- 
dence on the Lower Mississippi, and on a longer residence at the 
upper end of this great problem, among the mountains at the head- 
waters of the various tributaries of the Missouri, the following con- 
densed statement gives the results of this study, premising that it 
applies more especially to conditions as to the delta preceding the 
year 1861. During the "War between the States," the levees built 
previous to that date were destroyed to a greater or less extent, by 
military acts or by destructive floods. 

Each year there were two periods of flood-waters in the Mississippi 
River below the mouth of the Ohio. The first flood, styled the "spring 
rise," came almost entirely from the Ohio River, reinforced to some 
extent by the melted snows and rains from the Lower Missouri and 
the Upper Mississippi proper. The advent of this flood at its mouth 
varied in time, volume, and intensity with the climatic conditions of 
each spring. 

The second flood, styled the "June rise," came mostly from the 
Missouri River, caused, as a governing factor, by the melting of the 



60 THE FLOODS OF THE MISSISSIPPI DELTA 

accumulated snows of winter at the sources of its larger tributaries 
having their origin in the great Continental Divide. 

The advent of this flood at Cairo, and its volume and intensity, 
were dependent mostly on conditions at its head; the amount of snow- 
fall during the previous winter on its eastern water-shed, more 
especially in the vast pine forests; the time, whether early or late, of 
the first warm spells of spring, their duration and intensity (it re- 
quired a week or often ten days of warmth sufficient to melt snow at 
night, at an elevation of 8 000 ft., to produce the highest floods in the 
local streams) ; and, to some extent, on the local rains in the lower 
water-shed of the Missouri and Upper Mississippi Rivers. The spring 
floods from the Arkansas and Red Rivers, having their origin in the 
same range of mountains, but much farther south, did not contribute 
so seriously to the disastrous effects in the lower delta, because they 
usually came earlier, and had passed down before the advent of the 
"June rise" from the Missouri. 

In a majority of years the crest of the "spring rise" had passed 
Cairo before the advent of the "June rise" from the Missouri, in 
which event it passed down to the Gulf with but little damage to 
plantations or levees. In cases where cultivated fields were over- 
flowed, the water subsided in time to plant cotton and corn and to 
raise a partial crop of each. 

Following in the w^ake of the "spring rise" came the "June rise," 
but, as the former had substantially passed out of reach, the latter 
passed down to the Gulf without doing material damage to either 
levees or plantations. 

As is evident from this statement of conditions, if, from various 
climatic causes, the "June rise" from the Missouri came tumbling 
down upon the "spring rise" before the latter had passed out of the 
way, there would be one of those disastrous overflows which the system 
of levees built previous to that date had never withstood, resulting in 
enormous losses to the plantations, usually, in the entire delta. Such 
floods were so late in subsiding from the cultivated fields — sometimes 
as late as August 1st — as to prevent the raising of either cotton or 
corn for that season. The best information was that these disastrous 
floods came in cycles of about seven years, and varied in intensity, in 
accordance with the volume of water brought down in each of the two 
"rises," and, in a large measure, as to the point at or near Cairo, at 
which the two floods merged. 



THE FLOODS OF THE MISSISSIFPI DELTA 61 

Since the writer's personal knowledge of conditions in this delta 
up to 1873, more than thirty years ago, these conditions have been 
modified to a certain extent. The "spring rise" from the Ohio River 
water-shed makes its advent below Cairo earlier and in larger volume 
than in former years, in a volume so much larger that the levees are 
often crevassed with the sequence of great damage to the planting 
interests. Although this "spring rise" may have passed down and 
out of the way of the "June rise," yet, if the latter was of an average, 
or greater than the average, volume, its waters would overflow through 
the "crevasses" made by the earlier rise, widening them and resulting 
in about the same amount of damage as would have occurred had the 
conditions been the same as they were previous to 1861. 

Notwithstanding the somewhat earlier advent of the "spring rise" 
at Cairo, it has not prevented, at times, that conjunction of events 
which occurred in former years, when the two "rises" combined their 
floods there, with such disastrous effect in the delta below. The only 
perceptible difference is that these combinations of the two "rises" 
occur at somewhat longer intervals than seven years, as the writer 
is informed. 

The much earlier advent of the "spring rise" at Cairo can be 
attributed to but one cause, namely, the denuding of its entire water- 
shed of the timber and underbrush which covered it thirty or forty 
years ago. The result, as was to be expected, is that the melted snows 
and rains are precipitated into its valleys much earlier and in larger 
volume, resulting in great destruction of property, and in hardships 
to the people of Pittsburg and the entire Ohio Valley below. 

The destructive effects of the "spring rise" on the delta below are 
caused when the early floods from the Cumberland and Tennessee 
Rivers merge near its mouth with the floods from the upper tributaries 
of the Ohio. 

This condition of affairs in the Ohio Valley is an object lesson as 
to that which will occur on a much larger scale in the great delta, 
unless we not only preserve, but add to, the extensive forests now 
existing at the sources of the principal tributaries of the Missouri, 
which furnish such a large percentage of the annual floods poured 
into it at Cairo. 

Imagine for a moment that the forests on the eastern slope of the 
great Continental Divide could be suddenly wiped from the face of 
the mountain, and that their winter snows were exposed to the direct 



62 THE FLOODS OF THE MISSISSIPPI DELTA 

rays of the sun in spring and summer; there would be such an early 
melting of the snows in the spring that the "June rise" from the 
Missouri would make its advent near Cairo so early every year that 
it would merge with the "spring rise" from the Ohio, overtop the 
levee system, and create such devastation that in a few seasons it 
would depopulate the entire delta, and New Orleans would become a 
thing of the past. This is not a fancy sketch. The same causes, how- 
ever unrestrained, that have produced the conditions before alluded 
to on the Ohio Eiver water-shed (the greed of Man), combined with 
the ravages of forest fires, which, when under full headway are as 
uncontrollable as the floods of the Mississippi, render this catastrophe 
not impossible. Fortunately for the country, this is not a supposable 
case at this time. 

President Eoosevelt, soon after assuming office, aroused public 
opinion as to the value of these forests for irrigation, so that Congress 
quickly responded in liberal appropriations, the Forestry Service was 
at once organized along liberal lines and under the direction of 
Gifford Pinchot, Assoc. Am. Soc. C. E., and everything points to a 
thorough oversight of this most valuable asset. 

A brief study of the foregoing facts and views seems to point to 
a simple and evident solution of the problem set forth in the beginning 
of this paper, that is, to control the "spring rise" from the Ohio or the 
"June rise" from the Missouri so as to prevent their conjunction or 
merging near the mouth of the Ohio, which has always produced great 
destruction below. 

Could this desirable object be effected, a system of levees in the 
delta adequate to take care of the "spring rise" from the Ohio would 
substantially be sufficient to take care of the "June rise" which, 
though of greater volume, would flatten out when not encountering any 
back flow from the Ohio flood. 

There does not appear to be much relief in sight in reference to the 
control of the "spring rise" from the Ohio. Reforesting its water-shed 
would give relief to its immediate valley, but none below its mouth. 
On the contrary, the earlier advent of that flood at its mouth gives it 
the more time to get out of the way of the "June rise." The proposal 
to store the flood-waters of this stream in immense reservoirs, by 
impounding its waters behind dams, does not impress the writer as 
either feasible or practicable. The cross-section of its valley below 



THE FLOODS OF THE MISSISSIPPI DELTA 63 

Pittsburg is too flat for reservoirs of large capacity, and, above that 
point, the valleys of its tributaries are too narrov^^ and their gradients 
too steep to be available for that purpose. The same conditions exist 
as to its other important tributaries, the Cumberland and Tennessee 
Rivers. 

For the past ten years the General Government has been engaged 
in building an extensive system of reservoirs for the purpose of im- 
pounding the spring and summer floods at the head of the Mississippi 
River, with the object of turning them loose in the fall to assist in the 
navigation of this river belov^^ St. Paul. While these works answer 
and will continue to answer the purpose of their construction, the 
volume of water impounded will not give material relief at Cairo. 

It is, then, to the Missouri River that we must look for some means 
of holding back or retarding the advent of the "June rise," so as to 
give the "spring rise" sufficient time to get out of its way. 

The building of artificial dams to impound these floods seems to be 
neither feasible nor practicable, within reasonable cost, as was con- 
tended regarding the Ohio River floods. The writer recalls sites for 
such dams on most of the large tributaries of the Missouri (the upper 
ends of canons, with the valley of the stream above opening out into 
;i comparatively broad valley), where, by building dams 500 or 600 ft. 
high, much of the flood-water of each stream could be held back. With- 
out going into the calculations of the cost of such works on all the 
large tributaries, it would appear to be so enormous, when compared 
with the cost of the means suggested by Nature, and spread out all 
around these proposed dam sites, that comparison of the merits of each 
scheme would seem to be unnecessary. 

On the head-waters of its principal streams, Nature has pointed out 
the means to be used in accomplishing this desirable end by providing 
those dense pine forests that serve as vast storage reservoirs for the 
accumulated deep snows of winter. These forests, by their shade, pro- 
tect the snows from the direct rays of the sun, allowing them to melt 
so gradually as to prevent damage in the streams below. It is the rapid 
melting of the snow on the large areas on and contiguous to the 
Continental Divide, which from various causes have been denuded 
of their forests, that causes flood damage in the streams below, and 
is the main factor of the "June rise." 

The lead pointed out by Nature must be followed. The present 



64 THE FLOODS OF THE MISSISSIPPI DELTA 

area of these forests must not only be preserved, but must be added to, 
by reforesting the large areas of open or prairie land which are sur- 
rounded by or are contiguous to the main bodies of forested land. 

On each slope of the Continental Divide, in Montana and Wyoming, 
and on the mountain spurs branching out therefrom, are vast pine 
forests extending from 30 to 60 miles in width on each side of the 
summit. There are also groups of mountains of considerable extent 
arising from the plains on each side and within 100 miles of this 
Divide (such as the Big Horn Eange and the Teton groups), which 
are covered more or less with similar forests. 

The limit of the timber line in that latitude is about 9 200 ft. 
above sea level. Below that line and above 8 000 ft. elevation, and sur- 
rounded by or contiguous to these forests, are open or prairie lands 
about equal in area to that of the pine forests, which have been denuded 
of timber, by fire and other causes, the conditions of soil and moisture 
being similar to those of the dense forests adjoining them. 

On these upper mountain plateaus are to be found many never- 
failing springs, and these, during the dry months of the fall, will 
serve to irrigate the young pine shoots used in reforesting. The ad- 
joining timber will furnish all the small pine saplings needed. 

Above the timber line and on each side of the Continental Divide 
and the mountain spurs branching out therefrom are large areas of 
open land, but whether it has been denuded of its timber by fires, or 
whether its condition is normal and due to the frigidity of the climate 
in winter, is not apparent. On the eastern slopes there is generally a 
sufficient depth of good soil which has been drifted from the western 
slope by the high westerly winds prevailing at all times. 

During the storms of winter, there form on the eastern slope snow- 
drifts deep enough to last very often until snow flies again. The 
writer is not sufficiently informed as to the principles of forestry to 
give an opinion as to whether these open areas above the timber line 
can be reforested; but there are large areas with sufficient soil, and 
water from the snow banks, for that purpose, if the winter temperature 
does not preclude the idea. 

There are no records in existence as to the present extent of the 
pine forests on the head-waters of the Missouri River tributaries, nor 
of the open areas now denuded of forests from unknown causes, lying 
among or contiguous to the existing forests below the timber line. 



THE FLOODS OF THE MISSISSIPPI DELTA 65 

It is believed, however (and this belief is based on a somewhat 
intimate knowledge of about 100 miles length of this Continental 
Divide and its numerous spurs), that below the timber line there are 
areas of open land which can be reforested, and that these are sufB- 
cient to equal in extent the areas of existing forests. 

As before stated, the record of snowfall at the Four Bear Cattle 
Ranch, on the Grey Bull River, for a term of years, averaged about 80 
in., and its climate could be taken as the average of conditions nt 
points of similar elevation (7 000 ft.) around the water-shed of the 
Missouri River. It is thought that the snowfall on the high mountain 
plateaus is greater than in the valleys bejow, but there are no data at 
hand to sustain this opinion. 

The snowfall, as measured each day, was of light weight, the 80 in. 
of average snowfall representing about 8 in. of water. 

This 80-in. depth of snow has fallen between September 1st and 
May 1st, and, in the pine forests, at altitudes of from 7 500 to 9 200 ft., 
and represents all the rain and snowfall between these dates. At that 
altitude freezing temperature is one of the principal preservatives of 
snow, especially in the warm months. There are few nights at that 
season, among the pine forests, when snow does not "crust over" suffi- 
ciently to support the weight of a man or even a horse. 

In estimating the amount of precipitation that goes to make up the 
floods of the Missouri River, there must be added to this 80 in. of snow 
the precipitation in May, June, and July of each year. This represents 
the total annual precipitation for the year, except that which occurs 
in August. As the record for August around this water-shed is only 
about YTm in-j it can be safely assumed that this annual precipitation 
of that altitude represents the stored up precipitation for each year in 
those pine forest reservoirs. 

As will be hereafter explained, one-half of this precipitation (say, 
6 J in.) represents the portion covering the open areas, surrounded by 
or contiguous to the forests, which turns into water in May or during 
the first days of June, and flows into the larger tributaries within 12 
hours after melting, thus forming the controlling factor of the famous 
"June rise." The volume and intensity of this flood are caused, not 
only by the volume of water thrown into the stream within the space 
of a month, but by the melted snow being at once tumbled down the 



66 THE FLOODS OF THE MISSISSIPPI DELTA 

steep mountain into the water channels below in an incredibly short 
space of time, thus adding to the intensity of the flood. 

As contributory to the "June rise" may be mentioned the melting 
of the deep snowdrifts of the great plains west of the 100th meridian 
and the local rains of the Lower Missouri and Upper Mississippi. 

Contributory to the "spring rise," to some extent, are the melted 
snows of the open areas of these mountains, that lie above timber line 
(9 200 ft.). As these snows lie usually in immense drifts, they do not 
melt much before the latter part of July and August, too late to 
affect the "June rise." 

The volume and intensity of the "June rise" are dependent to a 
certain extent on the coincidence or conjunction of all these factors. 
The melted snows from the forest reservoirs are, undoubtedly, the 
controlling factor. 

In April, sometimes earlier, under the influence of the "Chinook" 
or warm winds from the Pacific, the snows covering the plains below 
the foothills of the Continental Divide commence melting, causing the 
first floods of the main stream and tributaries, which soon completely 
sweep out the ice, often creating immense gorges down as low as 
Omaha. 

The writer's information is that the influences of these Chinook 
winds at its head, frees the Missouri of ice much sooner than is the 
case with the Upper Mississippi. 

By April 1st, the snows in the timber on the high mountain 
plateaus have usually settled to a depth of from 4 to 5 ft. On 
account of the low temperatures peculiar to an altitude of 8 000 or 
9 000 ft., the snow does not usually commence melting in the open 
tracts until about May 1st. It melts so rapidly, however, that by 
June 10th the grass is in such an advanced stage that cattle from 
the ranches below are taken, at about that date, to these upper moun- 
tain plateaus to be kept there until the November snows. 

In the forests contiguous to these open spaces, the snow does not 
usually disappear until the beginning of August. 

This, in a nutshell, tells the value of forests in conserving the 
snows of winter. In other words, forests of average density and at 
the average altitude of their habitat, 8 000 to 9 000 ft., will preserve 
the snows of winter from melting two months longer than open areas 
exposed to the direct rays of the sun. 



THE FLOODS OF THE MISSISSIPPI DELTA 67 

As an object lesson in support of this experience on the Grey Bull, 
the writer cites his experience on July 16th, 1880, in crossing the Con- 
tinental Divide from the waters of Snake River opposite the Upper 
Geyser Basin of the Yellowstone National Park. The Divide is com- 
paratively flat, has an elevation of about 8 200 ft., with about an aver- 
age density of forests, and at that date was covered with a depth of 
from 3 to 6 ft. of snow, depending somewhat on its exposure to sun- 
shine at the small openings in the timber. The area of these openings 
ranged from a few acres to 20 acres. There may have been some drifting 
of snow in places. As is always the case at this altitude, the snow 
banks are crusted over during the night from cold (one of the elements 
of snow preservation at this altitude), and the party was compelled 
to take advantage of it, and get the pack animals across the snow- 
drifts on the summit early in the morning. The small openings in 
the forests, before alluded to, had been freed from snow so long that 
grass was so far advanced in growth as to afford enough pickings for 
nooning for the horses. The previous winter had been an average one 
for snowfall, and that summer an average for heat. This experience 
afforded an admirable object lesson as to the efficacy of forests for 
preserving snow. 

Now apply these facts and conditions to the entire Missouri River 
water-shed of like altitude, 7 000 ft. and above, to the Milk, the Sun, 
the Jefferson, the Madison, the Gallatin, the Yellowstone, the Big 
Horn and the Platte Rivers. As at present, one-half of the winter's 
snows that cover the open spaces will pass off in May or thereabouts, 
and will not be of benefit for local irrigation on the smaller streams, 
for it is not then needed, nor on the larger streams below for naviga- 
tion, for local rains have already provided a STifficiency. 

Now conceive the entire water-shed of this mighty river (below 
the timber line and well above the agricultural belt, say 7 700 ft.) 
to be clothed with forests of average density. When spring comes, the 
snow will commence melting in that belt about May 1st, and instead 
of half the winter's snowfall melting during that month, and causing 
as a main factor the "June rise," it will melt gradually, and will not 
have disappeared before August 15th. Such gradual melting will not 
create those spasmodic floods, common in the local streams below, that 
often do so much damage to irrigation work. The floods in all the large 
tributaries are held back by these causes, and, in the outcome, the 



68 THE FLOODS OF THE MISSISSIPPI DELTA 

"June rise," about which so much has been said, will reach Cairo on 
an average one month later than usual. If the levees of the delta 
are of average strength, they will have taken care of the "spring rise" 
without any serious damage having been done. It will have passed 
down and out of the way of the "June rise," held back from causes 
before described, and the latter rise, from the same causes, will be of 
much less intensity than in former years, and will pass down to the 
Gulf and do but little damage. 

The "June rise" is not caused entirely by melted snows, but is re- 
inforced to some extent by local rains on the Lower Missouri and the 
Mississippi. The retarding of the melting snows of the "June rise" 
one month, by the means before described, will result in lessening its 
intensity to that amount, as the effects of the local rains will have 
passed down. 

Reforesting the upper water-shed of the Missouri, as pointed out, is 
a work more expensive in the time required in its accomplishment than 
its cost in dollars and cents. 

No plan has been suggested for the control of the floods of this 
mighty river which will not require years of time, patience, and labor. 
It is believed that the plans laid out by the Mississippi River Com- 
mission, as understood, are along conservative lines, as regards this 
great delta, that is, to hold the river to its present channel by the 
revetment of its concave or caving banks, by contracting its channel 
at bars by proper works, and, where necessary, by dredging these bars 
for navigation purposes. In addition to these works is Considered (as 
the writer understands the views of the Commission) a substantially 
built system of levees on each bank. To carry out these plans will 
require much time and patience, with proportionately much labor and 
expense. 

To make a permanent success of these works in the delta, it is 
considered essential that the high floods of the Ohio and Missouri 
water-sheds shall be kept apart. The only feasible means of accom- 
plishing that desirable object is by impounding the flood-waters of the 
latter stream in the immense forest reservoirs at its head. 

Reforesting the open areas among these forests, will be small in 
expense compared to the work in the delta. For a few years it will 
require the labor of as many men as can be worked to advantage for 
five months of each year, in resetting the pine saplings, and in making 



THE FLOODS OF THE MISSISSIPPI DELTA G9 

the small ditches necessary for irrigating them. This will require 
about 1 000 men for a few years. After the shoots have taken root, 
fewer men, who should be expert at irrigation, will be required. 
Nature, with the warm rays of the sun, will do the rest. It is believed 
that within five years after this work is started its beneficial effects 
will be apparent. 

To make a success of this reforesting will require the application 
of the laws of forestry suitable for each latitude of the North American 
Continent and to each species of pine found most suitable; and this 
must be combined with the experience of expert irrigators and much 
patience and care. 

When it is considered that the successful accomplishment of this 
reforesting is one of the main factors, in conjunction with the con- 
templated works in the delta below, in adding to the resources of the 
Nation each year from $300 000 000 to $400 000 000, as estimated here- 
tofore in a detailed report, it becomes a matter of national importance, 
and should be carried through regardless of whatever expense, within 
reasonable bounds, its accomplishment may require. 

Closely allied to, and as one of the most important outgrowths of, 
forest preservation is the Reclamation Service, brought to the front by 
the broad-minded statesmanship of President Roosevelt and at once 
adopted by Congress. Its design is to reclaim, under a wisely guarded 
law, the arid regions of the mountain States on each slope of the 
Continental Divide, by expending the proceeds of the sales of the 
public lands in the construction of reservoirs and ditches leading there- 
from for purposes of irrigation. One of the most direct benefits to the 
Nation from the construction of this reclamation system, when finished, 
will be the large annual saving in the production of the sugar beet, 
now so quickly and profitably changed into sugar, one of the neces- 
saries of life. The belt of country lying between the 100th meridian 
and the Continental Divide and between New Mexico and the 
Canadian line, is eminently adapted to the production of the sugar 
beet. It is believed that the portion of this belt, under the influence 
of the Reclamation Service, on each slope of the Great Divide, when 
brought under proper cultivation, can produce enough sugar, in 
addition to what is now being produced, to take the place of the 
2 500 000 tons now imported from foreign countries. 

In other words, these millions of dollars now being spent on the 



70 THE FLOODS OF THE MISSISSIPPI DELTA 

Reclamation Service will add each year to the resources of the Nation, 
when completed, at least $200 000 000, besides other benefits of great 
value which will naturally suggest themselves. 

Bearing on the question of its influence in holding back the flood- 
waters of the streams making up the great "June rise," A. P. Davis, 
M. Am. Soc. C. E., Chief Engineer of the Reclamation Service, has 
kindly furnished the following table of the "Irrigating Projects" under 
construction in the Missouri River drainage basin : 

Projects. 

Huntley, Mont 

Sun River, Mont 

North Platte, Nebr 

Lower Yellowstone 

Milk River, Mont 

Williston, N. Dak 

Belle Fourche, S. Dak.... 
Shoshone, Wyo 



Acreage. 


Discharge, in 
second-feet. 


storage, in 
acre-feet. 


33 000 


400 




300 000 


3 500 


600 000 


300 000 


3 500 


1 000 000 


67 000 


800 




80 000 


1000 




40 000 


500 




100 000 


1200 


200 000 


150 000 


1800 


400 000 


1 070 000 


12 700 


2 200 000 



The experience derived from irrigation on a small scale on the 
local streams of that water-shed is that the water which goes through 
the ditches and is turned out from them, is soaked up by the soil and 
does not return to the stream for a month or more. This happens in 
eases where the ditches are not more than ^ mile from the stream. 
In the case of the above-named projects the main canals will be from 
1 to 3 miles distant from the streams from which they flow. It can be 
estimated safely that the water passing out of the main canals of the 
above "projects" during May, June, and July, does not return to the 
streams from which they radiate for such a length of time that this 
discharge, together with the water impounded by the dams, can be 
considered as reducing the volume of the "June rise" to that extent. 
The volume of water thus held back amounts, by calculation, to about 
4 500 000 acre-ft. ; not a large volume, yet there are times when it might 
have a perceptible efl'ect. 

It must be borne in mind that a .test has never yet been made as 
to the ability of the two earthen walls, comprising the levee system on 
each bank of the Mississippi between points opposite Cairo and to and 



THE FLOODS OF THE MISSISSIPPI DELTA 71 

below New Orleans, to withstand the pressure that would be brought 
to bear, did the "spring rise" from the Ohio and the "June rise" from 
the Missouri combine near Cairo and attempt the passage to the Gulf. 
Judging the future by the past, the result, with the existing system 
of levees, would at least be problematical. 

Neither the magnitude of the problem of the successful control of 
this river nor the benefits to be derived therefrom, are fully realized 
by the country. The history of engineering affords no precedent to 
serve as a guide in solving the problem. At flood tide its forces are 
irresistible when brought in direct antagonism to the puny work of 
Man. At times its actions appear almost as erratic as a child's. It 
has no respect for State's rights. In a few weeks time it will slice off 
a large area from one State and as quickly deposit it below on the 
opposite bank, to add to the territory of a neighboring State. It is 
only by humoring its eccentricities and guiding its inherent power, 
that Man can accomplish the solution of its control. This plan was 
successfully carried through at its mouth, when it was forced to dig out 
a channel 35 ft. deep (where a depth of 15 ft. had existed) across sand- 
bars formed by its own currents, by the construction of the jetties. 

The cost of the work necessary for the control of the flood-waters 
will be great. The benefit to the United States arising therefrom by 
adding to its resources 4 000 000 bales of cotton and a proportionate 
amount of sugar each year makes it evident that its cost should be 
borne by the Nation and not by the comparatively few cotton and 
sugar planters of the delta, as has heretofore been done through State 
agencies. It should be evident that it is as much the province of 
the General Government to take care of the harbors of this vast inland 
sea and the channels leading to them as to take care of the harbors 
and channels of the seaboard and the Lake States. 

As a means of keeping open this important channel of commerce, 
the building of levees, the revetment of the caving banks on their 
front, the contraction of its channel at low water in the interests of 
navigation, are the same means rendered necessary in the construction 
of the same works for the purpose of prevention of disastrous over- 
flows. Any plan of river improvement for navigation without the aid 
of levees on the banks, would be a serious mistake and would result in 
failure. A bad crevasse in the levees would decrease the volume of 
water below, and counteract one of the governing ideas in the plans of 



72 THE FLOODS OF THE MISSISSIPPI DELTA 

the Mississippi River Commission, viz., the contraction of the currents 
for scouring purposes. 

The foregoing points out an additional reason why the cost of this 
great work should be borne by the Government. It carries with it the 
result that this entire system of work — the control of overflows by 
levees, and the works for the improvement of navigation — shall be 
under one central control. The two systems go hand-in-hand, and are 
interdependent. 

To recapitulate: The solution of this great problem consists in 
keeping apart the "June rise" and the "spring rise," so that the latter 
floods will have passed before the advent of the former at the mouth 
of the Ohio. 

As noted before, there is no relief to be expected from the Ohio. 
It is left, then, to the Missouri water-shed to furnish the means for the 
object required. The head- waters of this stream must be impounded 
in immense reservoirs, for such a length of time as will without doubt 
prevent the "June rise" from making its advent until the "spring rise" 
has passed. 

Nature has constructed, as an object lesson, those immense storage 
reservoirs for protecting the winter's snows in the dense forests in and 
near the great Continental Divide. It directs us to reforest those 
large areas, which at one time were clothed with dense forests, but 
which have been denuded by fire and other unknown causes. 

Nature teaches us that these open spaces, once clothed with forests, 
can be reclothed with similar forests if Man will use the means indi- 
cated by Science. Then will follow the slow melting of the conserved 
winter snows, thus keeping apart the "June rise" and the "spring rise," 
and, in conjunction with a substantial system of levees in the delta, 
this will afford the only solution for the permanent control of the 
periodical floods that sweep down through this great delta. 

In the preparation of this paper, the writer is indebted to H. N. 
Pharr and W. I. Hardee, Members, Am. Soc. C. E., for information as 
to conditions in the Mississippi Delta subsequent to 1S73, and to A. P. 
Davis, M. Am. Soc. C. E., of the Interior Department, and to Messrs. 
Kellogg, Bonsteel, Plummer, and Zoo, of the Agricultural Department, 
for valuable statistics. 



AMEEIOAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1102 



ELECTKIC RAILWAYS IN THE OHIO VALLEY 

BETWEEN STEUBENVILLE, OHIO, AND 

VANPOKT, PENNSYLVANIA.* 

By George B. Francis, M. Am. Soc. C. E. 



With Discussion by Messrs. F. Lavis, George B. Preston, J. Martin 
ScHREiBER, William J. Boucher, and George B. Francis. 



There has been constructed during the past two years, along the 
northerly and westerly banks of the Ohio River, between Vanport, Pa., 
and Steubenville, Ohio, about 40 miles of double-track, standard-gauge, 
electric railroad, which affords this busy, thriving, industrial section a 
high-grade interurban road. 

The lines form the connecting links uniting the East Liverpool, 
Wellsville, Steubenville, and Toronto systems, and, with their con- 
nections, at Vanport on the north and Steubenville on the south, makn 
possible through travel by trolley between Pittsburg, Pa., and Wheeling, 
W. Va., except for a short piece of track, now under construction, in 
the vicinity of Sewickley. 

The configuration of the country through which the road has been 
built is such as practically to preclude the location and construction 
of a future competing line. 

* Presented at the meeting of January 6th, 1909. 



74 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

The railways described have been built by three constituent 
companies : 

The Steuhenville and East Liverpool Railway and Light Company. 
—This company owns and operates the properties of the Steubenville 
Traction and Light Company and the Toronto Electric Light and Power 
Company. The former of these acquired properties owned and operated 
the street railway system in Steubenville, and a single-track line to 
Toronto on the north, a distance of about 10 miles. The new company has 
reconstructed this single-track road, improving its alignment and grade 
at various places by locating the tracks on a private right of way; has 
added a new second track between Steubenville and Toronto, and has 
built a new double-track road between Toronto and Wellsville, a dis- 
tance of 7.86 miles. The company also furnishes light and power for 
commercial purposes in Steubenville and Toronto. At Steubenville, 
the southerly terminus of the system, connection is made with the 
Wheeling Electric Railway, which reaches Wheeling, W. Va., a dis- 
tance of about 21 miles, also with the Steubenville and Brilliant Line, 
on the west side of the river. 

The East Liverpool Traction and Light Company. — This company 
serves the East Liverpool district, which extends eastward as far as 
the State line between Ohio and Pennsylvania, and southward to the 
southerly limits of Wellsville, and includes the street railway system 
in East Liverpool, a branch line across the river to Chester, W. Va., 
and a 3-mile spur track to the company's coal mine up Island Run. 
It also furnishes light and power for commercial purposes in East 
Liverpool, Chester, and Wellsville. The initial move in improving and 
increasing the transportation facilities demanded by this territory was 
made by this company. The road was originally a single-track line, 
lying wholly in the highways and streets, and had several dangerous 
steam railroad crossings at grade. In the reconstruction of that por- 
tion of the system which forms a part of the through main line, pro- 
vision was made for a double track, which has been laid partly in new 
location, and as far as practicable on private right of way, improving 
the grade and alignment, and eliminating the grade crossings. 

The major portion of this property was .reconstructed during 1906 
and 1906, thus preceding the construction of the stretches of road 
southward to Toronto and Steubenville, and northward from East 
Liverpool to Vanport, which has resulted in the continuous line of 
about 40i miles of double-track road described herein. 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 



75 




RAILWAY LINES 
STEUBENVILLE, O. TO ROCHESTER, PA. 



Not to Scale 



Existing Lines Reconstructed and Doiible-Tiackcd 
■"« New Lines Constructed of Double Track 



DISTANCE8 ON THROUGH LINE. 

''Steubenville & E.LivcrpooI Railway & Light Co. 18.29 Mile 

East Liverpool Traction & Ligbt Co. 10.83 

Ohio River Passenger Railway 11.29 '■ 

Beaver Valley Tr. Co.- Vanport to Rochester 1.00 " 

Total 41.41 " 



Fig. 1. 



76 ELECTKIC RAILWAYS IN THE OHIO VALLEY 

The Ohio River Passenger Railway Company. — This company has 
built the new double-track road between East Liverpool (where con- 
nection is made with the lines of the East Liverpool Traction and 
Light Company) and Vanport, a distance of 11.29 miles. At the 
latter place connection is made with the lines of the Beaver Valley 
Traction Company, with which company a traffic agreement has been 
made for operating cars into Rochester, Pa. Arrangements are being 
perfected with the Pittsburg and Lake Erie Railroad for a connection 
at Beaver and for the sale of tickets for through travel to any point on 
either system. 

A tri-party agreement has been executed with the foregoing con- 
stituent companies permitting the joint use of the main-line trackage 
between Steubenville and Rochester, a distance of 44J miles. 

Location.— The portion of the Ohio Valley traversed by these con- 
necting lines is narrow, with high and, in some places, steep and 
precipitous hillsides and bluffs alternately advancing and receding 
from the shore of the river. 

The location along the bank of the river on a low grade line was 
practically pre-empted by the Cleveland and Pittsburg Division of the 
Pennsylvania Lines, and it required considerable engineering study 
and skill to secure private right of way, and locate a line which would 
not only give easy curves and grades, but avoid expensive trestles and 
viaducts. 

Railway Construction. — The permanent way consists of a graded 
roadbed, located partly in the streets and highways, but for the greater 
portion of the distance on a private right of way, laid with 85-lb. 
rails, and ballasted in a first-class manner. The alignment and grades 
are such as are compatible with high-grade urban and interurban roads. 
All streams and waterways are crossed by substantial bridges and 
culverts, the former being of steel and the latter of concrete or stone 
masonry. 

TracJc. — The track is of standard-gauge construction, and in the 
main is of 85-lb. T-rail, Am. Soc. C. E. section, laid in 60-ft. lengths 
on 6 by 8-in. by 8-ft. hardwood ties, mostly oak, spaced 24 in. from 
center to center, and ballasted with gravel or broken stone. The rails 
are bonded for the return circuit with single 000 protected bonds, 
except in East Liverpool, where double bonds of the same type are used 
on each joint. Cross-bonds connect the rails at intervals of 1 000 ft. 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 



77 



W K K ft K 





§ 


u 


y 




i 


\ 


10 

s 




B 






i: 


jtt 


III 


ei 


w 


11 


, 






















to 




^ 


f- 


■ 












^ 


fi 


Vi 


tunti 


■e 
11 


'a 


•k 




F 
















CO 


( 


i' 


^ 




















\ 


<, 


5- 














05 






> 
















Of 




/ 


\ 
















S 


N 

4 


■% 
















.„iC 


oston 


la 








S r„ 




^ 


j4 












< 


i 




^ 


ar 


tie 










i - 








1 


01 


Ol 


tc 








C 

5' 




1 


/ 


















Cio 


;tonl 


luii 


1 




1 


laiis 


1 






/ ^ 


Impii 


1 
e 










ts 


\p 
















-1 




;k( 


y> 


mi 

1 










ts 




















8 




< 


•<c 


p 


>rt 
C 
G 


Homer 

& p.k. 

oofe R„ 


I. 

n 








<j 


Brin 


St 


>D< 


i 


UI 
























^ 


'% 












7^ 


I 


rellov 


■ ( 


re 


!k 










j 










ts 




\ 


Bor 


.*P 


Cross 


ngs 

1 






^K^ 










1 


^i 




(1 












' 






P- 
















K M t?3 HE tS 



O Ol 

o o 



78 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 



The tracks are generally 9 ft. 8 J in. from center to center; but 
where bridges or trestles are crossed, provision is made for spacing 
the tracks 10 or 12 ft. from center to center. Spring frogs of standard 
pattern are used in all cases outside of city streets. Fig. 3 shows the 
standard adopted for roadway construction, outside of towns and cities, 
both in cuts and fills. 

Grading. — Outside of cities and towns the grading was practically 
side-hill work, necessitating some heavy cuts and fills. In many in- 
stances the railroad right of way parallels the highway which was 
originally cut in the sides of these precipitous hills. This roadway 
was widened, providing room for both the highway and the double- 
track railway line. 




^ STANDARD ROADBED SECTION 

TIES 6-IN.X 8-IN. 8 FT.LONG 
RAILS A. S.C.E. SEC. WEIGHT 85 LB.PER YD..60-FT. LENGTHS. 
GRAVEL BALLAST 

Fig. 3. 

The work required the excavation of about 750 000 cu. yd. of 
njaterial of all classes, a considerable percentage being rock, mostly 
shale and sandstone. 

Beginning at Seventh Street, in Steubenville, the principal work 
may be described, in sequence toward Vanport, as follows: Grading 
and paving in highway for track and street widening for 2.76 miles; 
thence grading and culverts for double track on private right of way 
across the King and other property for 2.23 miles, t\iis embracing con- 
siderable earth and rockwork; thence track grading and highway 
paving and widening for 1.42 miles; thence earthwork grading and 
bridging on private right of way through the Toronto Realty Com- 
pany's property for 0.30 mile to the streets of Toronto ; thence grading 
for track in the streets of that town for 2.2 miles; thence on private 
right of way and over a double-track, timber trestle at Croxton's Run, 
360 ft. in length, for 0.40 mile; thence grading for track, partly in 



PLATE XIII. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1102. 

FRANCIS ON 

ELECTRIC RAILWAYS IN THE OHIO VALLEY. 




Fig. ].— Retaining Wall, Cooks Ferry, Ohio. 




Fig. '.i.— Yellow Creek Bridge. 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 79 

streets, but mostly on private right of way through Calumet, Empire, 
and Ekeyville for 4.07 miles to the north end of the steel viaduct, at 
Port Homer, over the Cleveland and Pittsburg Railroad, 450 ft. long; 
thence side-hill grading on private right of way, including relocation 
of highways along what is known as The Narrows, to Yellow Creek, 
for 3.01 miles, this stretch including steel viaducts at Goose Run, 
335 ft. long, Brimstone Run, 175 ft. long, various cattle passes, heavy 
retaining walls, and small bridges ; thence by a single-span steel bridge, 
174 ft. long, over Yellow Creek; thence under the Cleveland and 
Pittsburg Railroad through existing bridge openings, and side-hill 
grading, including the straightening and paving of the highway for 
0.78 mile to Borings Crossing, a new bridge being there constructed 
for the Cleveland and Pittsburg Railroad over the new road and high- 
way, thus removing an existing grade crossing; thence track grading 
and street paving in Wellsville, for 3.69 miles to the city line of East 
Liverpool, this stretch including the Wellsville steel viaduct, about 
300 ft. long; thence track grading and street paving in East Liverpool 
for 7.14 miles to the easterly city line, this stretch including the 
Jethro Run steel viaduct, about 475 ft. long, the Sixth Street steel 
viaduct, 664 ft. long, Bradys Run culvert, Dry Run bridge, and other 
structures. For a considerable portion of the latter distance the high- 
ways were widened, and some of the new railroad was placed on a 
private right of way; thence crossing Beaver Creek by a two-span 
double-track steel bridge, 306 ft. long, and crossing under the Cleve- 
land and Pittsburg Railroad through an existing opening, and in 
streets for 1.47 miles to a bridge constructed over the Cleveland and 
Pittsburg Railroad tracks, for both the new railroad and the highway, 
thus removing a grade crossing; thence track grading on a private 
right of way, and straightening, widening, and paving the highway 
for 2.20 miles to Midland; thence track grading and paving for 1.51 
miles through this town; thence grading on a private right of way for 
1.17 miles to Industry; thence track grading and street paving through 
this village for 0.37 mile; thence grading on private right of way 
(mostly side-hill work) for 4.55 miles to the end of the new double 
track at Vanport. This last stretch embraces the following steel viaducts 
and other important structures : Four Mile Run viaduct, about 323 ft. 
long; Barclay Run viaduct, about 275 ft. long, and Vanport trestle, 
about 700 ft. long. 



80 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

The side-hill grading was quite difficult, as a great number of slides 
came down the hillsides, in some instances from considerable dis- 
tances, causing large expense for their removal. The clayey nature of 
the material, combined with water from the hills, was the cause of 
many of these slides. The excavated material consisted of earth, a 
mixture of earth and rock fragments of varying sizes, and solid rock. 

The side-hill work necessitated the construction of a number of 
retaining walls at different points, the principal ones being at The 
Narrows, just south of Yellow Creek, and one near Cooks Ferry. The 
former is about 350 ft. long and has a maximum height of 30 ft.; 
each contains approximately 1 600 cu. yd. of concrete. The wall at 
Cooks Ferry is of similar din "nsions. 

Paving. — The line passes through fifteen communities, and in 
nearly every instance where a franchise was given for the use of 
streets or highways unusual requirements for paving were exacted. 
In some instances the entire width of the street was paved; in others, 
where the street grade was changed, the embankment slopes in front of 
residences were paved, as well as the street surface. 

As this is a great brick manufacturing center, all paving consisted 
of vitrified brick on a sand cushion, 11 miles of track being thus paved, 
the cost averaging approximately $0.90 per sq. yd. 

Bridges. — All important bridges, with the exception of those over 
Beaver and Yellow Creeks, are deck structures of the viaduct type, 
weighing about 700 lb. per lin. ft. of double track, and were calculated 
for a concentrated load of 24 tons on two axles at 10-ft. centers, or a 
uniformly distributed load on trusses of 1 800 lb. per lin. ft. of each 
track. 

The standard of construction calls for 35-ft. braced towers with 
battered posts, and 60-ft. spans between the towers, with provision for 
expansion at each tower. All columns and towers are fixed to the 
foundations with long anchor-bolts set in the piers, and at the abut- 
ments the structure is held in place by short bolts set after the erection 
of the steel, thus allowing for creeping during erection. The struc- 
tures, though light, were built for intern rban electric passenger service 
only, and are thoroughly braced and very rigid. The floor systems are 
stiffened further by carrying the ties clear across the structure. All 
foundations and abutments are of concrete. Fig. 4 is a typical design 
for these viaducts. 



PLATE XIV. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1102. 

FRANCIS ON 

ELECTRIC RAILWAYS IN THE OHIO VALLEY. 




fir.. 1, — Tk,\ck View at Indistrv. Pa. 




Fiii. iJ.— Water-Cooling Tower. Steubenville Power Plant. 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 81 

The bridges over Beaver and Yellow Creeks are double-track, 
through, tvpo-truss structures, having rigid members with riveted 
connections. The load is distributed evenly over the bearings by using 
pin-connected shoes. At the expansion end the shoe rests on a nest 
of turned rollers which are free to move in a planed roller box bolted 
to the masonry. The structures are secured firmly to the concrete 
abutments and piers by long anchor-bolts set in the masonry. 

The Beaver Creek bridge consists of two double-track, through 
spans, each 153 ft. long; the Yellow Creek bridge consists of one double- 
track, through span, 174 ft. long. The steel for these bridges was 
fabricated and erected by the Fort Pitt Bridge Works. 

At Sixth Street, East Liverpool, a steel viaduct was built over 
what is locally known as Horn Switch. It is 664 ft. 4 in. over all, 
is 40 ft. above the foundation at its highest point, and weighs about 
2.33 tons. The steel for this bridge was fabricated by the American 
Bridge Company, and erected by the Cleveland Engineering Company. 

The aggregate weight of all the new truss bridges and viaducts 
approximates 1 100 tons. 

A number of small I-beam bridges were built, and three of the 
existing highway structures were strengthened to take care of the 
loading imposed by the addition of a second track. 

Power Requirements. — The power requirements of the original 
companies forming a part of this property were provided for in four 
generating stations, located at East Liverpool, Wellsville, Toronto, and 
Steubenville. These stations supplied current, not only for the street 
railway operations, but also for municipal lighting and power. To this 
load factor were added the requirements for the operation of about 
60 miles of new track and an increase of 50% in the lighting loads 
at East Liverpool, Steubenville, and elsewhere. 

In the changes, alterations, and additions to the generating, dis- 
tributing, and transmitting system, the existing houses and equipment 
were utilized as far as possible. 

After a careful study of the conditions, it was decided to consoli- 
date all the generating equipment in the East Liverpool and Steuben- 
ville power-houses. The loads formerly carried by the Toronto and 
Wellsville stations were transferred to these main stations, and the 
buildings were remodeled and equipped as sub-stations, one new sub- 
station being built at Industry. 



82 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

The difficulties of this work were increased by the necessity of 
keeping the generating stations in continuous operation while the 
changes and alterations were being effected. The work was accom- 
plished successfully without the interruption of any service that was 
being rendered. 

East Liverpool Power-Eouse. — The original building was a brick 
structure with a wooden floor and a tar and gravel roof on timber 
supports. The old wooden floor was taken out and replaced with rein- 
forced concrete, giving increased cleanliness and greatly reduced 
fire risk. 

The load on this station was increased by the addition of that 
formerly carried by the Wellsville station and the new Industry sub- 
station, as well as by the increased requirements of the railroad, to 
provide for which the alternating-current generating equipment was 
increased by the addition of one 1 000-kw. turbo-generator and one 
500-kw. turbo-generator, and the direct-current equipment by one 
new 300-kw. rotary converter and the transfer of a 300-kw. engine- 
type generator from the Steubenville power-house. A new 7-panel 
switch-board was installed to take care of this additional electrical 
apparatus. 

The turbines as well as the alternating-current, belt-driven genera- 
tors at this plant produce 60-cycle, 2-phase current at 2 200 volts. 
To reduce this voltage for the operation of the rotary converter, two 
l75-kw. step-down transformers were installed, and to raise it for 
transmission to Wellsville and Industry, six l75-kw. step-up trans- 
formers were provided. Excitation is taken care of by one steam- and 
one motor-driven exciter, either of which is of sufficient capacity to 
furnish excitation for all the alternating-current generators. A 
Tirrill regulator was installed for voltage regulation. 

The boiler plant was enlarged by the addition of a wooden frame 
structure, sheathed with corrugated iron, in which were installed four 
500-h.p. Stirling water-tube boilers. A 2 000-h.p., open feed-water 
heater, and two 12 by 7 by 12-in. boiler feed pumps were also 
provided. 

In order to operate the turbines, a barometric condenser with dry 
vacuum pump and two motor-driven circulating pumps were installed. 
Water for condensing purposes is taken from the Ohio Kiver, and, as 
there is about 40 ft. difference between the high and low stages of the 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 



83 




84 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

river, unusual means had to be provided for elevating the water to the 
condensers. To this end a vertical intake or well of reinforced concrete 
was built on the bank of the river, extending from below the low- 
water to above the high-water level. Vertical centrifugal pumps were 
then installed at the bottom of this well and connected by vertical 
shafts to two 50-h.p. 2 200-volt induction motors placed above the high- 
water mark. From the bottom of the head-house a 24-in. intake pipe 
was extended to the center of the channel of the river, terminating in 
a timber crib. 

The coal supply for this plant is obtained from the company's mine, 
from which it is carried over a 3-mile branch line and delivered in 
dump cars at the fire doors of the boilers. The generating capacity 
of the station has been increased from 1 200 to 3 000 kw. 

Steuhenville Power-House. — This is a brick structure with a brick 
floor and a slate roof, supported by steel trusses. A separate building, 
of similar construction, was enlarged to provide for the installation of 
two 500-h.p. boilers. 

The station load was augmented considerably by the transfer of the 
lighting load formerly carried at the Toronto station, the increased 
lighting load for Steubenville, and the additional power required for 
the operation of the 25 miles of newly constructed railway lines which 
it serves. To take care of these requirements, the following additional 
equipment was installed : Two 500-kw. turbo-generators, producing 
60-cycle, 2-phase current at 2 200 volts ; two 175-kw. and two 100-kw. 
step-up transformers for raising the voltage to 13 200 volts, 3-phase. 
for transmission to the Toronto sub-station. The switch-board was 
enlarged by the addition of four new panels to handle this new equip- 
ment. Excitation is taken care of by one steam-driven and one motor- 
driven exciter, either of which is of sufficient capacity to furnish 
excitation for all the alternating-current generators. A Tirrill regu- 
lator was also installed. 

The plant was enlarged further by the addition of two new 500-h.p. 
Stirling boilers, and the necessary condensing apparatus for the tur- 
bines, consisting of a barometric condenser, dry vacuum pump, and 
engine-driven centrifugal and circulating pumps. 

As the only source of water supply for condensing purposes was 
the city water system, it was necessary to construct a cooling tower 
and an additional engine-driven centrifugal pump to deliver water 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 85 

thereto. In erecting this tower, use was made of an old gas tank and 
well on the property. 

The capacity of this station has been increased from 1 800 to 
2 500 kw. 

Toronto Sub-Station. — The Toronto sub-station building formerly 
contained the generating apparatus supplying current for the arc- and 
incandescent-lighting systems of Toronto, and was built of tile blocks, 
with a brick floor and a slate roof. As now equipped, it contains two 
300-kw. rotary converters furnishing direct railway current at 685 
volts, and operated from the transmission line through six 110-kw. 
step-down transformers, and two 110-kw. transformers which deliver 
2-phase current at 2 200 volts for the incandescent-lighting circuits, and 
for the arc-lighting circuits by constant-current transformers and mer- 
cury arc rectifiers. The station has a 10-panel switch-board. 

Current is received from the Steubenville power-station at 13 200 
volts, over two 3-phase transmission lines 9^ miles long. 

Wellsville Sub-Station. — The apparatus for this sub-station is in- 
stalled in a building which was formerly the Wellsville power-station, 
supplying Wellsville with arc lights for street lighting and incan- 
descent lights for commercial purposes, as well as direct current for 
street railway operation. 

The building is of brick, with a solid concrete floor and a tin roof. 
Current is received at 6 600 volts from the East Liverpool power-house, 
over two 3-phase transmission lines each 4| miles long. 

The present sub-station apparatus is designed to take care of the 
commercial and railway loads, and consists of two 300-kw. rotary 
converters furnishing current at 585 volts for railway purposes, the 
current being delivered to the rotaries from the transmission line 
through six 110-kw. step-down transformers; and four 75-kw. trans- 
formers which deliver 2-phase current at 2 200 volts for the incan- 
descent-lighting circuit. The arc lights are taken care of by two con- 
stant-current transformers supplied by the 2 200-volt bus. The station 
also contains a 200-kw. "balancer set," furnishing direct current at 
500 volts for a special commercial circuit, supplying shop motors which 
could not be operated at the railway voltage. The station has a 9-panel 
switch-board. 

Industry Suh-Station.- — This sub-station is a new brick structure, 
with a concrete floor, and a reinforced concrete roof supported by 



86 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

I-beams. It contains two ^UU-kw. rotary converters which receive 
current from the transmission line through six 110-kw. step-down 
transformers and dehver direct current at 585 volts for railway service 
only. Current is transmitted to this station at 13 200 volts from the 
East Liverpool station over a single 3-phase transmission line about 
10 miles long. A 5-panel switch-board is installed. 

Overhead Construciion. — The usual type of span-wire construction 
used by direct-current interurban roads was adopted. The poles are 
of chestnut, from 30 to 35 ft. long, with 7-in. tops. They are set 6 ft. 
in the ground, with head and breast boards, and are given a deflection 
away from the tracks of 18 in. in 24 ft. The poles are approximately 
100 ft. apart, with a minimum clearance of 24 ft. from center to 
center of poles, measured at the track level. The poles on one side 
of the road are 5 ft. higher than on the other, and carry, in addition 
to the trolley, the feeder and transmission hues. The butts of all 
poles carrying high-tension wires are charred for a distance of about 
18 in. above and below the ground line. The tops and gains of all 
poles were painted with two coats of heavy black asphaltum before 
erection and before the attachment of the cross-arms. The cross-arms 
are painted yellow pine, fitted with locust pins, secured to the poles 
by through galvanized-iron bolts, and braced with galvanized-steel 
braces. 

The trolley lines are mainly of 00 hard-drawn copper wire sup- 
ported by I -in. galvanized stranded span wire to which they are 
attached by round-top hanger suspensions with 15-in. bronze soldered 
ears, except on curves, where the ears are carried by gooseneck sus- 
pensions with wood strain insulators. They have a clear height above 
the track of 19 ft. 6 in. outside of city limits and on private right of 
way, while in the cities this clearance is increased to 21 ft. 6 in. Pro- 
tection from lightning is furnished by arresters placed on the poles at 
intervals of about 1 000 ft. 

Transmission and Feeder Lines. — The six wires forming the double 
3-phase transmission lines between East Liverpool and Wellsville are 
supported on one 8-ft. cross-arm, the wires of each circuit being 16 in. 
apart. The double transmission lines between Steubenville and 
Toronto are carried in delta on two cross-arms, with the apex of the 
triangle at the bottom, the cross-arms being spaced so that the wires 
form a 24-in. equilateral triangle. Between East Liverpool and Tn- 



ELECTRIC RAILWAYS IN THE OHIO VALLEY 87 

dustry, the single 3-phase transmission line is carried on a 7-ft. cross- 
arm, the wires being spaced 24 in. apart. This construction makes 
possible, with the least alteration, future provision for a double 3-phase 
line, similar to the Steubenville and Toronto lines, by simply adding 
another cross-arm with two insulators. 

The wires are carried on brown porcelain, high-tension insulators, 
51 in. in diameter and 4^ in. in height, and are protected from light- 
ning by suitable lightning arresters and choke-coils in each of the 
power-houses and sub-stations. The lines consist of No. 2 B. «& S. hard- 
drawn copper wire, except between East Liverpool and Wellsville, where 
No. 4 wire is used. This wire is bare, except through built-up districts, 
where it is protected by water-proof insulation. At two points on the 
route the transmission lines are carried on independent pole lines which 
have been erected for short distances on a private right of way. 

In addition to the direct-current railway feeders required for local 
service, the main line of the road is taken care of by 795 000-cir. mil. 
aluminum feeders carried on a 2-pin cross-arm below the transmission 
line, taps to the trolley being provided at intervals of 1 000 ft. 

Cars. — New rolling stock has been purchased for the through service 
between Steubenville and Rochester, and consists of 18 double-truck 
Pullmans, 44 ft. long, finished in mahogany, with high-backed, red 
plush seats, having a seating capacity of forty-four. Cars of this type, 
and for service of this character, are usually fitted for greater carrying 
capacity, but in this equipment four seats, two on either side, were 
omitted, the seats being respaced to give more liberal distance between 
them, following more nearly, in this particular, steam railroad practice. 

Each car is equipped with four 60-h.p. motors, mounted on Brill 
trucks with steel wheels. A smoking compartment, with a seating 
capacity of sixteen, is provided at one end; all cars are heated and 
lighted by electricity, and fitted with electric arc headlights, Lecturn 
signal lights, electric bells, parcel racks, and other conveniences. 

The exterior finish is a rich yellow, lettered in aluminum outlined 
in black. On the letter-board over the windows, the cars bear the words 
"Ohio Valley Scenic Route." 

Organization in the Field for Roadway Construction. — The engi- 
neering of those portions of the roadway of the East Liverpool Traction 
and Light Company's property constructed and reconstructed during 
1905 and 1906 w^s looked after by Westinghouse, Church. Kerr and 



88 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

Company. The work was executed under various contracts at unit 
prices by others and by the company's own forces. 

The engineering of the subsequent roadway extensions of this com- 
pany's property, the construction and reconstruction on the Steuben- 
ville and East Liverpool Railway and Lighting Company's property, 
and the work on the Ohio River Passenger Railway Company's prop- 
erty constructed during 1907 and 1908 was looked after and constructed 
directly by the forces of Westinghouse, Church, Kerr and Company, 
as was also the power work during 1907 and 1908. 

The roadway organization in the field consisted of the engineer in 
charge, assisted by field parties and superintendents on the different 
main divisions, with sub-divisions under general foremen, walking 
bosses, foremen of gangs, etc., these being revised from time to time 
as the work demanded in its several parts, such as grading, masonry, 
track-laying, paving, and overhead work. Some unavoidable difficulties 
in obtaining right of way delayed the work so that the time covered 
was more than would have been necessary otherwise. The maximum 
number of men employed at one time was about 1 000. No steam 
shovels were used or required. Four locomotives, with about 30 dump 
cars, were utilized more or less continuously in the grading and in the 
distribution of track material and ballast. There were various concrete 
mixers, derricks, and pile drivers. The total cost of the roadway work 
for 1907 and 1908 was about $1 600 000. The cost of the power work 
done during 1907 and 1908 aggregated about $450 000. 

Government Work. — Extensive river improvements are being made 
by the United States Government in the construction of a deep-water 
channel between Pittsburg and Cairo and the building of dams and 
locks which will make navigation possible on the Ohio throughout the 
year. It is estimated that at the present time, with these improvements 
still far from completion, more than 10 000 000 tons of freight per 
annum are moved in boats and barges during the high-water stages, 
over the 40 miles of river between Beaver and Steubenville. 

Population. — 'Table 1 gives the approximate population of the vari- 
ous cities, towns, and districts through which the road passes, or which, 
by reason of connections made, may fairly be considered as contribu- 
tory; also the approximate mileage. 

Industries. — The section of the Ohio River ■ between Beaver and 
Steubenville is rich in resources — coal, natural gas, gil, clay, quarries, 



ELECTRIC RAILWAYS IX THE OHIO VALLEY 89 

etc.— which make it a natural industrial center. Here are situated 
some of the large iron, steel, and tin-plate works of the United States 
Steel Corporation and a number of independent companies; extensive 
plants for the manufaature of vitrified sewer pipe, fire brick, paving 
brick, tile and other fire-proofing materials, as well as glass-works and 
paper mills. ^^^^ ^^ 



District. 


Mileage. 


Population. 


Pittsburg 

Intermediate 

Beaver 

liiterniediat? 


8 
17 

4 
11 
11 
10 

8 
19 
U 

102 


700 000 

.%000 

46 000 

8200 


East Liverpool 

Intermediate 


44 500 
18 900 


Steubeuville 

Intermediate 

Wlieeling 


41500 

11500 

100 000 




1000 000 



East Liverpool is the principal seat of the pottery industry in the 
United States. Three of the twenty-seven potteries located here are 
said to be the largest in the country; and this industry, which alone 
employs 12 000 hands, with a weekly payroll of more than $100 000, 
produces more semi-porcelain tableware, door knobs, and porcelain 
electric fixtures, than any other city in the world. 

Other important manufacturing interests, related or semi-related 
to the foregoing, make this 40 miles of river valley "a veritable hive of 
industry." 

Amusement Park. — One of the most beautiful and popular amuse- 
ment resorts in this section of the country is Rock Spring Pjirk, at 
Chester, W. Va., just across the river from East Liverpool, and reached 
by one of the company's branch lines. It constitutes the main 
pleasure resort in this part of the valley, and no expense has been 
spared to maintain the high standard of its attractions. It is owned 
by the East Liverpool Traction and Light Company, and is operated 
by an amusement company to which it is leased. 

Scenic Attraction. — The road passes through an exceptionally pic- 
turesque country, and gives promise of excellent possibilities for the 
development of a large excursion business; it will undoubtedly prove 
an attraction to pleasure seekers. 

For its entire length, the route follows the river, and is located 
generally on the table lands affording many fine views of the 0-hi-yu 



90 ELECTRIC RAILWAYS IN THE OHIO VALLEY 

or "Beautiful River," which fully justifies the name given to it by the 
Indian tribes formerly inhabiting its banks. 

In the vicinity of Toronto, great apple orchards cover the hiUs for 
almost 10 miles, and in the spring the great mass of white and pink 
bloom against the fresh green of the hillside presents a strikingly 
beautiful picture. The beginning of the apple industry is traced to 
the eccentric woodsman, John Chapman, familiarly known as "Johnny 
Appleseed," who, during the latter part of the eighteenth century, 
devoted his life to planting apple trees in this vicinity. 

Operation. — The various sections of the road were placed in opera- 
tion as fast as completed, as follows : 

Double track between Wellsville and East Liverpool, Sept. 15th,1906. 

Between Calumet and Port Homer, Dec, 14th, 1907. 
*Between Steubenville and East Liverpool, Feb. 11th, 1908. 

Between East Liverpool and Smith's Ferry, Feb. 29th, 1908. 

Between Smith's Ferry and Midland, March 21st, 1908. 

Between Steubenville and Calumet, June 30th, 1908. 

The entire road between Steubenville and Vanport, July 1st, 1908. 

Cars are run on 30-min. schedule on the main through line, with 
extras and specials as required. 

Financiers, Engineers, and Constructors. — The construction of these 
railway lines has been financed by the Ohio Valley Finance Company, 
the Hon. W. Caryl Ely, of Buffalo, N. Y., President. The general 
supervision of the whole project, the securing of right of way, and the 
purchase of rolling stock was looked after by Mr. Edward McDonnell, 
of East Liverpool, Ohio, Assistant Treasurer, and Mr. Van Horn Ely, 
of Steubenville, Ohio, Vice-President of that company. 

Westinghouse, Church, Kerr and Company, of New York City, were 
the engineers of construction and reconstruction of the old properties, 
and the engineers and constructors of the new properties, the road- 
way construction being under the supervision and direction of the 
writer assisted by Edwin J. Beugler, M. Am. Soc. C. E. The execu- 
tion of the work in the field was under the direction of William V. 
Polleys, M. Am. Soc. C. E., Resident Engineer. Mr. George B. 
Preston was the engineer in charge of the mechanical and electrical 
work in the field and office. 

*A part of the distance between Steubenville and Calumet was operated as a single-track 
road until June 3Uth, 1908. 



DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLHY 91 

DISCUSSION. 



F. LaviSj M. Am. Soc. C. E. — The valley of the Ohio River and Mr. Lavis. 
the territory tributary to it have witnessed a greater development of 
long-distance interurban electric railroads than any other part of the 
United States. Many of these lines run through sleeping, dining and 
drawing-room cars, built by the Pullman Company, of only slightly 
lighter construction than the standard equipment of the largest trunk 
lines of steam railroads. A description, therefore, of one of these 
lines, completed only about six months ago by one of the most prominent 
engineering firms identified with work of this kind, is presumably a 
description of the very latest and best practice in the construction of 
railroads of this type at the present time. 

The general characteristics of these interurban lines are, that they 
pass through the various cities and towns along the route on the sur- 
face of the streets, and as near the business centers as possible, while 
in the open country between, they are located, for the greater part of 
the distance, on private right of way, in many instances having 
numerous highway crossings at grade. This latter feature is not men- 
tioned by the author, and may have been avoided on this road, but the 
number of these crossings, which are only slightly, if at all, less 
dangerous than those of steam railroads, is becoming increasingly 
large, even in States where a great deal of public money has been, and 
is being, expended in eliminating such features on steam railroad lines. 

This road, between Rochester and Steubenville, passes through a 
territory described by the author as a very "hive of industry," and the 
population tributary to it, amounting to more than a million people, 
or practically 10 000 per mile on tlie whole length between Pittsburg 
and Wheeling, is composed largely of the class of prosperous mechanics 
and laborers which furnishes a larger patronage, in proportion to the 
total population, to a road of this kind than does any other. Tlie most 
substantial construction, therefore, is apparently warranted. 

In order to get some idea of the relative importance of the con- 
struction on such a line, it seems desirable to have some basis of com- 
parison with other standard railroad practice. The bridges on this line 
are designed for a concentrated load of 24 tons on two axles at 10-ft. 
centers, or 1 800 lb. per lin. ft. of single track. The heaviest equip- 
ment of the New York Subways and Elevated Lines has a load of 
30 tons on the two axles with the motors, and 11 tons on the other 
two, or an average of 2 080 lb. per lin. ft. of track, which is only 
slightly greater than that used on this line. The largest Pullman cars 
in use on steam railroads are about 80 ft. long, and weigh about 60 
tons, this load being carried on two six-wheeled trucks, giving about 
10 tons per axle and about 1 500 lb. per lin. ft. of track. A passenger 
locomotive of the Atlantic type, weighing about 75 tons, which is a fairly 



92 DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 

Mr. Lavis. large engine for express passenger service on anything but mountain 
grades, has a load of about 20 tons per axle on the driving wheels, and 
the load, for engine and tender loaded, will be about 4 000 lb. per lin. 
ft. of track. A consolidation freight engine, weighing about 100 tons, 
has a load of about 22 tons on the driving axles, and will average 
5 000 lb. per lin. ft. of track. The bridges on this road, therefore, will 
carry almost any standard railroad equipment except locomotives, the 
weights of the heavier types of which in ordinary use would be about 
double that for which these bridges were designed. 

Having in view, therefore, the fact that the requirements of roads 
of this class are not, in many respects, very far removed from standard 
steam railroad practice on first-class lines, and are in some respects 
greater than the requirements on many inferior lines, it has seemed to 
the speaker that it would greatly enhance the value of the paper if 
the author would give some of the reasons for the very radical 
departures from steam railroad practice as regards the location of 
this line. 

The author states that "It required considerable engineering study 
and skill * * * to locate a line * * * which would give easy 
curves and grades," considering the natural difiiculties of topography, 
and the possibility of securing private right of way. Nothing is 
stated in the paper in regard to what the author considers to be easy 
curves for a road of this type, although, as a matter of fact, outside of 
the cities, curves of from 200 to 300 ft. radius, provided there were not 
too many of them, and the amount of central angle covered was 
necessary, would probably have little effect either on cost of operation 
or speed. In passing through cities, curves of as small a radius as 
50 ft. are not infrequent. This difference between roads of this type 
and trunk-line railroads where a curve of 1 000 ft. radius is con- 
sidered quite sharp and necessitates slowing down to from 20 to 25 
miles an hour, is quite marked, and is due of course to the absence 
on the electric road of the long rigid wheel base of the steam 
locomotive. 

The question of grades on these electric roads, however, is one 
about which far less is known, at any rate publicly. Eailroads build- 
ing new lines to be operated by steam, even in a great deal of the, as 
yet, undeveloped parts of the country, are very reluctant to adopt higher 
rates than 0.6% unless the topography of the country absolutely 
demands it, whereas on the profile of the road under consideration, as 
shown by Fig. 2, one grade of 9.9% is shown for a fairly long stretch, 
and there are several stretches of 6% to 7%, although most of them 
seem to be fairly short. 

The reason, of course, for the badly broken grade line is the 
fact that the river bank is occupied by the steam railroad, thus forcing 
the electric road in many cases high up on the side-hills where these 



DISCUSSION ON KLECTRIC RAILWAYS IN THE OHIO VALLEY 93 

come close to the river and necessitating its descent nearly to the river Mr. Lavis. 
level where the hills lie farther back, in order to avoid a long detour 
to keep up. This characteristic of the country is shown very clearly 
in the view of the Yellow Creek Bridge. 

The ability to ignore, to a very large extent, considerations which 
affect the determination of the rates of grade, which govern steam 
railroad practice, is the chief characteristic which differentiates the 
location of interurban electric railroads, or any electrically operated 
railroad where single cars or multiple-unit trains are to be used 
exclusively, from that of those operated by steam or electric locomotives 
at the head of the train, although few data have as yet been made 
public as to the actual conditions which govern the determination of 
the most economical grades to be used on these electric roads. 

On a steam railroad, in conjunction with the necessity of low 
grades, it is also almost as important to arrange the grade line so that 
the demand on the locomotive will be as nearly even as possible 
throughout comparatively long stretches. No considerations of this 
kind apparently influenced the fixing of the grades on this railroad, 
and here again it would undoubtedly be of considerable interest to 
know just what considerations governed the judgment of the engineers 
in adopting the grades they did on this particular road. 

As the steam locomotive is most economically operated when the 
demand on the machine is fairly consistent, so also is this the case 
with almost any machine or power plant, including, and perhaps 
especially so. electrical power plants. 

In the case of the latter, when power is being manufactured for 
the operation of a railroad, it may be considered that, where a fairly 
large number of cars are being operated, the demand will be equalized 
by the fact that some cars will be going down grade while others are 
going up, some stopped while others are going at full speed, etc. ; but, 
on a road such as the author describes, where the service is infrequent 
and the number of cars in use at a time small, the very broken grade 
line and the high rates of grade would apparently tend to create a 
somewhat irregular demand, as there would be nothing to prevent a 
combination frequently occurring where all the cars, or at least a 
majority of them, would be going up hill at once, and then within a 
few minutes be all going down hill. In the case of power-houses 
supplying power only for the operation of the cars on a railroad of 
this kind, this 7night be a serious consideration. The power-houses on 
this line, however, supply power for both manufacturing purposes and 
electric lighting, for which purposes the load is probably fairly even, 
and is a large proportion of the total output. The variation in the 
load from the operation of the cars, therefore, probably does not affect 
them greatly. The great economy of a uniform load at a power-station 
is strikingly exemplified by the contract recently made by the Com- 



94 DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 

Mr. Lavis. monwealth Edison Company, of Chicago, for supplying power to 
the street railway companies of that city.* The very low rate of 0.4 
cent per kw-hr. was made, provided the demand was uniform, any 
variation from this uniform demand being penalized according to a 
certain schedule mutually agreed upon. Of course, a street railway 
operating in any large city will have high peak loads during the rush 
hours, and therefore the demand cannot be regular; this instance, how- 
ever, being so recent, and the contract so important, serves to emphasize 
the desirability of uniform loads at the power-houses. 

It is well to bear this in mind and not lose sight of the fact that the 
electrical operation of railways has its limitations, although many 
liberties may be taken with the grades and alignment that would not 
be permissible on steam railroads. 

There is another question which it has seemed to the speaker might 
naturally occur to some of the readers of the paper, and that is: 
Before deciding on the construction of these lines at all, was any 
consideration given to the possible effect on the new line, should the 
steam railroad, with its truly easy grades and curves, decide to 
electrify its line and r\m short trains at frequent intervals, or even 
equip its line with one of the many types of motor cars now apparently 
in successful operation on many steam railroad lines, where frequent 
service for a comparatively small number of passengers at a time is 
demanded? Probably the steam railroad would be at a disadvantage 
by its inability to reach the business centers of the various towns, and 
also to leave every passenger nearly at his own front door. It is quite 
probable, also, that such a service as that necessary to compete success- 
fully with the electric lines might interfere too much with the through 
business of the road, both passenger and freight. 

In constructing an interurban railroad across country, some years 
ago, where 60-ft. rails were used, the speaker had some trouble from 
the excessive expansion and contraction in such comparatively long 
lengths, until the ballast was filled in as close as possible to the top 
of the rails and on both sides of them. If the temperature changes 
are at all large, and the joints are left open sufficiently to prevent the 
track from kinking in hot weather, there is likely to be excessive pound- 
ing of the ends of the rails at the joints, due to the wide spacing. It 
would be interesting to know whether any such trouble has been found 
on this road; and, if so, what means were taken to overcome it. The 
standard cross-section of the track, shown in Fig. 3, seems to show 
that no ballast was placed above the tops of the ties. 

It is quite common practice now, in laying street railway tracks 

through paved streets, to lay a fairly substantial concrete foundation 

both under and around the ties, and to such a height over them as 

will just allow for the type of paving to be used, will) the necessary 

*'En(iineerin(] Newn, December 10th, IfiOs. 



DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 95 

sand cushion under it. The speaker had occasion to note only a short Mr. Lavis. 
time ago the installation of this type of construction for a street rail- 
way track operated by a large public service corporation, on a double- 
track line, where the cars were operated on 20 min. headway, with little 
if any special excursion business, so that the traffic was probably not 
greater than on the line described by Mr. Francis. No mention of 
such a type of track construction is made in the paper. If it was not 
used through the paved streets in the cities, was it on account of the 
necessity of curtailing expenses, or because it was not considered 
economical, in view of the expected traffic on the road? The excessive 
wear and tear on the motors of electric cars due to poor track has led to 
the very general adoption of the very best type of track construction 
possible for street railroads almost everywhere. 

The width between centers of tracks, adopted on this road, namely, 
9 ft. 8J in., is the general practice on street railway work in cities, 
though many interurban roads, where high speed is used, have adopted 
a wider spacing through the open country. On the road recently com- 
pleted between Baltimore and Washington the tracks were laid 11 ft. 
on centers, and 80-lb. rails in 33-ft. lengths were used. The fact that 
the bridges on the Ohio road were designed to allow the tracks to be 
spaced 12 ft. on centers allows the inference that the engineers had in 
view the possible necessity of wider spacing of the tracks in the future. 

There is another point which the speaker believes is not made quite 
clear in the paper, and that is the reason for building a double-track 
road. The only information as to the amount of traffic is the state- 
ment that cars are operated on 30 min. headway for ordinary traffic, 
and that there is a considerable excursion business. On roads with 
cars operating normally on 30 min. headway, it is generally quite 
satisfactory to build a single-track road with sufficient sidings to pro- 
vide for operation on half that headway, the speaker having in mind a 
road in New England which handles quite successfully a very large 
summer excursion business on this basis. 

The construction of steam railroads is now carried out along line? 
v.'hich have become fairly well standardized, and the type of construc- 
tion necessary to meet certain conditions is in a general way well 
known to engineers at all familiar with this branch of the profession. 
The construction of electric railroads, however, owing to the much 
greater flexibility of the application of power, is subject to much 
greater variations, and a most intimate co-relation of the work of the 
engineers responsible for the location of the line and the electrical 
engineers is necessary in order to produce the most economical results. 
The practice in the construction of this road, as described by Mr. 
Francis, is so radically different from steam railroad practice that the 
speaker believes that a fuller statement of the reasons which led to the 
final design, so to speak, of this road, would be most interesting, not 



96 DISCUSSION ON ELECTRIC EAILWAYS IN THE OHIO VALLEY 

Mr. Lavis. Only to the members of this Society, but to all engineers interested in 
work of this class. 
Mr. Preston. George B. Preston, Esq. — As having some bearing on the relation 
between the average and maximum loads on the power stations, it may 
- be of interest to note that with the heavy holiday load of July 4th, 
1908, the 1-hour maximum at the East Liverpool house was 2 000 kw., 
while the average load during the heaviest part of the day, from 
Y to 11 P. M., was 1 800 kw. On an ordinary day the 24-hour average 
for this station is 850 kw. 

When considering the question of peak loads, in connection with 
this system, it should be borne in mind that these two power-houses are 
furnishing all the commercial and city lighting, as well as carrying a 
considerable commercial motor load, for four cities having an aggre- 
gate population of 73 000. 

Mr. Boucher has asked whether or not the possibility of using the 
third-rail was considered in connection with this railway system. It 
was not considered feasible to use the third-rail system because of the 
necessity of utilizing the local electric railways at Steubenville and 
East Liverpool which were already installed and equipped with the 
overhead trolley, and also from the fact that a considerable portion of 
the roadbed lies in public streets and highways. 
Mr. schreiber. J, Martin Schreiber, Assoc. M. Am. Soc. C. E. (by letter). — 
Although no new engineering features, other than represented in 
modern practice, are broiight out in the paper, it is an intelligent and 
comprehensive description of a very interesting electric railway. 

It is gratifying, at least to those who have had experience in the 
maintenance of some of the old-time electric railways, to know that the 
capitalist is rapidly realizing that it even pays to spend some money 
on the permanent improvement of street railways. Indeed, the very 
near future will see almost all electric roads constructed with as much 
skill and forethought as is represented on steam roads. This will 
be necessary for economical, satisfactory, and profitable operation, and 
in order to compete with other carrying companies. 

The physical difficulties encountered in building the track and road- 
way of the Ohio Valley lines were out of the ordinary, as the location 
was not favorable for the work, and the author seems to be justified 
in the assertion that the configuration of the country through which 
the road was built is such as practically to preclude the location and 
construction of a future competing line. No doubt this condition 
accounts for the variable and at some places severe grades of the track. 
However, since the road is double tracked, successfid and satisfactory 
operation is facilitated. 

The author states that 85-lb. rails, in 60-ft. lengths, and of the 
Am. Soc. C. E. section, were used throughout the construction and 



DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 97 

reconstruction. The writer is of the opinion that rails in 60-ft. lengths Mr. Schreiber. 

are not altogether desirable on open railway work. It is true that there 

are fewer joints and more of a continuous track with a longer rail, but 

the^te is a disadvantage in handling them, and difficulty in maintaining 

proper alignment. Unless special expansion joints are used, 60-ft. rails 

will kink in summer and have open joints in v.inter. The 3.3-ft. rail seems 

to work to the best advantage for open track, and the 60 and 62-ft. 

lengths for city or paved track. 

From one of the photographs it appears that the rail in paved 
city streets is also of the T-section. The advisability of using the 
T-rnil in paved streets and adopting it as a standard in all street 
railway work has been strongly advocated by many railroad men. The 
Way Committee of the American Street and Interurban Engineering 
Association will make this subject its principal work during 1909, and 
it is hoped that a definite recommendation will be offered at its next 
annual convention. 

A number of authorities, including the engineers of cities, are co- 
operating with the street railways in permitting the installation of 
T-rails in the streets. Several of the up-State New York cities, includ- 
ing TJtica and Syracuse, laid T-rails during 1908, and they are 
reported as very satisfactory; and such cities as Minneapolis, with 
T-rails and brick paving, claim to have some of the finest street-railway 
track and roadway. 

The great advantage of the T-rail, over any other section, from an 
operating standpoint, and especially with the increasing wheel loads, 
is too well known to explain at length, and the question is often only 
in regard to permission from the authorities for its installation, as 
they generally object on the ground that the rail does not allow proper 
paving facilities, and that the paving wears too • rapidly along the 
gauge line of the rail. 

It is indeed unfortunate that in the early installation of T-rails in 
city streets, the paving, especially adjoining the rail, was not properly 
executed. This placed the T-rail in disrepute for municipal work, so 
that it is not uncommon for the engineer of a small borough to demand 
the installation of a Trilby rail, similar to that required in Philadelphia 
or New York. The writer has often fotmd this to be the case where 
vehicular traffic was practically nil, and where the paving was macadam. 
Here the T-rail would not only be cheaper and a great deal more 
desirable by the railroad company than any other of the common 
girder sections, but would give better service and more satisfaction to 
the public at large. This paving about such a rail is as easily main- 
tained as that adjoining the Trilby rail, and with the Trilby section 
the broken stone is continually getting into the groove and breaking 
wheel flanges, causing uncomfortable riding, disagreeable noises, and 



98 DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 

Mr. Schreiber. frequent derailments, not to mention the trouble of maintaining the 
■ line and gauge. 

On account of the great improvement made in T-rail track con- 
struction, it is the consensus of opinion of railway engineers that it is 
only a question of time when municipal authorities will co-operate 
with railway companies in the adoption of the T-rail, except in the 
very largest cities, where the streets are narrow, and traffic is very 
heavy. Just how far the T-rail may be used in densely populated 
cities, and with heavy wagon traffic and narrow streets, where trucks 
are compelled to follow the tracks, is difficult to anticipate; but, for 
ordinary city and interurban railways the T-rail should certainly 
displace the girder section. 

The great difficulty of maintaining track and keeping it in line, 
and with proper joints, with the Trilby and tram sections, are objec- 
tions which must be overcome; and there are railway managements 
which, in all probability, will direct their efforts in the future to satis- 
fying the municipal authorities, by building better track, with sub- 
stantial foundation, proper drainage, and more improved pavements 
to suit the T-rail, rather than attempt to operate with an unsatisfactory 
rail that will meet the present paving conditions. 
Mr. Boucher. WiLLiAM J. BoucHER, Assoc. M. Am. Soc. C. E. — This paper is 
interesting in several particulars, and draws one's attention to the very 
extended development of cross-country electric roads competing with 
steam roads in Ohio and Indiana, many of which publish time-tables 
and schedules, run "limited" trains, and traverse distances from 50 to 
100 miles on one division. As these roads are primarily to connect 
centers of population, they necessarily run through sparsely settled 
districts, and frequently on private right of way; the query then 
arises, why are so few roads operated by the "third-rail"? 

In any consideration of the cost of installation, it is necessary to 
specify the kind of work wanted. Single-track, span-wire trolley con- 
struction may be had at a variety of costs, but, for similar operating 
conditions and equivalent current capacities, there is a marked dif- 
ference in the cost of overhead and third-rail construction. Assuming, 
then, that first-class construction in both cases is called for; also 
similar operating and current conditions; a double-pole, single over- 
head trolley with feed wire; other material and labor necessary for 
erection may be had at a cost of from $4 600 to $.5 000 per mile, de- 
pending considerably on the nature of the ground in which the poles 
are placed. An equivalent third-rail system, using an SO-lb. rail on 
extended ties and insulators, with splice-plates, bonds, and under- 
ground cable for road crossings, will cost, erected, about $3 800 per 
mile of single track. Tlie third-rail is unprotected. Board protection, 
similar to that installed in the New York Subway, will cost about 
$2 000 per mile. The maintenance charges should be considerably leas 
for the third-rail system over a period of years. 



DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 99 

Noticing, then, this difference of $800 to $1 200 per mile in favor Mr. Boucher, 
of the third-rail system, there must he some compelling reasons why 
the overhead trolley is used so much more frequently. Snow should 
not cause trouble to a third-rail if the interval between cars is no 
greater than 30 min., and during a storm any road must be prepared 
to run all night in order to keep the road open. Sleet, however, may 
cause trouble, but a scraper has been devised and is in use in New 
York which seems to overcome the difficulty. Of course, in the case 
of roads operating through country and towns, the frequent shifting 
of shoe and trolley would prove a nuisance, especially if a double- 
throw switch must also be operated, and without the latter the shoe 
is dangerous to passengers who congregate around the steps and near 
the trucks while waiting to board the cars. In the case of long- 
distance roads, however, with expanses of country and few towns, and 
running alongside a platform at stations, the third-rail would seem 
to be superior. If the road on its private right of way is properly 
fenced, and cattle guards are used at crossings, but little trouble will 
be caused by cattle and horses. 

The following roads are now operating with the third-rail : Albany 
and Hudson; Wilkes-Barre and Hazleton; Lackawanna and Wyoming 
Valley; Aurora, Elgin and Chicago; Scioto Valley, Ohio. 

George B. Francis, M. Am. Soc. C. E. (by letter). — Referring to Mr. Francis. 
the discussion by Mr. Lavis: There were a few grade crossings created 
on the private right of way portion of the railways, but these could not 
be avoided with reasonable economy, and their creation was not 
prohibited by law or the public authorities. 

The railways described are of the mixed type, comprising electric 
city street railway and interurban railway, for passenger service only. 
These conditions oblige and warrant radical departures in lines and 
grades from steam railroad practice, even on the private right of way 
portions. These portions, however, were located so as to conform sub- 
stantially to steam railroad practice, taking into consideration char- 
acter of power, equipment, and service. Speeds on lines of this 
character, operated with the type of wheel flanges required in usual 
stroft railway service, are necessarily less than would be the case with 
master car builder's steam railroad wheel flanges. 

The steep grades indicated on the profile occur in places where 
the line is in public streets, and are governed thereby. Private right of 
way location was out of the question in such places, for many reasons, 
mainly because it was desirable to occupy certain main streets in best 
serving the public and increasing the revenue. 

Consideration was given to the possibility of the parallel steam 
railroad improving its passenger facilities by electrification or other- 
wise, but, for various reasons, the probability of their so doing was 
thought to be too remote to govern the construction or non-construction 
of the new line; the main reason being the interference of local 



100 DISCUSSION ON ELECTRIC RAILWAYS IN THE OHIO VALLEY 

Mr. Francis, passenger service with the requirements for freight service, in this 
particular locahty, on the steam railroad lines. 

No trouble has been experienced from expansion of the 60-ft. rail 
length adopted on this road. Such troubles, however, have occurred 
in situations of peculiar character elsewhere. 

No concrete foundations were used for either the track or street 
paving involved. The local street vehicular traffic did not seem to 
require the placing of such foundations under the paving, the brick 
paving generally used in this locality, laid on a sand or gravel ballast, 
appearing to retain a good surface after many years' use. 

The road was double tracked for several reasons, viz., safety in 
operation, adherence to schedule, adaptability to variable schedules 
(such as are demanded by holidays or changing conditions of service), 
and because such was the strong desire of the management controlling 
the properties. 

By referring to Engineering News of December 10th, 1908, it will 
be seen that Mr. Lavis' statement : "the contract recently made by 
the Commonwealth Edison Company, of Chicago, for supplying power 
to the street railway companies of that city," providing for a rate of 
0.4 cent per kw-hr., conveys a wrong impression (probably vminten- 
tional on the part of Mr. Lavis), and that an additional payment of 
$16 per annum per kw. of maximum demand is provided for in that 
contract; and this brings the cost to about 0.74 cent per kw-hr. The 
steady load making this figure possible is not the railway load, but 
the general lighting load of the Edison Company, which equalizes the 
railway peaks. 

Mr. Preston, in his discussion, appears to have covered the point 
about the third-rail raised in Mr. Boucher's discussion. 



AMERICAN SOCIETY OF CIVIL ENGINEERS 

1 N S r 1 T U T E I ) 18 5 -' 



TRANSACTIONS 



Paper No. 1103 

NICKEL STEEL FOR BRIDGES. 

By J. A. L. Waddell, M. Am. Soc. C. E. 



With Discussion by Messrs. Charles Evan Fowler, M. F. BRO^VN, 
II. P. Bell, L. J. Le Conte, W. K. Hatt, John C. Ostrup, T. 

ClaXTON FlDLER, ROBERT E. JoHNSTON, AlBERT LuCIUS, G. LiNDEN- 

thal, Henry S. Prichard, Henry Le Chatelier, A. Ross, L. 
Dumas, Victor Prittik Perry, W. H. Warren, Willum R. Web- 
ster, William H. Breithaupt, E. A. Stone, C. Codron, W. W. K. 
Sparrow, B. J. Lambert, William Marriott, Henry Rohwer, 
Samuel Tobus Wagner, A. W. Carpenter, Leon S. Moisseiff, 
James C. Hallsted, F. Arnodin, Wilson Worsdell, William F. 
Pettjgrew, and J. A. L. Waddell. 



In November, 1903, the writer began a series of experiments, upon 
the comparative values of nickel steel and carbon steel for bridge build- 
ing, for the purpose of preparing from them an exhaustive economic 
study of the subject of "Nickel Steel for Bridges." These experi- 
ments extended over three years. 

The first step was to collect and read all the literature upon nickel 
steel bearing upon the subject, and, as this was rather meager, the 
task was soon completed. Although much was learned in this way 
concerning nickel steel in general, little was ascertained about the 
composition of the alloy most suitable for bridge building. 

* This paper, as originally prepared, was much longer and more elaborate than as 
herein given; but, for the sake of ecoTiomy of space, its dimensions have been redueed in 
every possible way that did not injure its continuity or militate materially against its 
ultiniate object. The original paper with its six appendices is to be found in the Library of 
the Society, where those who desire any of the detailed information it contains and this 
paper doei luit, may refer to it. As it gives every step in the development of the investiga- 
tion, and -allows hoW the writer's ideas and intentions were modified from time to time as 
the inquii'y jmiceeded. it may prove of interest to those who are contemplating making 
important engineering investigations. 



102 NICKEL STEEL FOR BRIDGES 

At the outset, the writer retained the Osborn Engineering Com- 
pany, of Cleveland, Ohio, to conduct the experiments, under his im- 
mediate direction and supervision; and F. C. Osborn, M. Am. Soc. 
C. E., the President of that company, at first gave his personal atten- 
tion to the work, but, as soon as it was laid out, George C. Saunders, 
Assoc. M. Am. Soc. C. E., was placed in charge for the company. 

In order to ascertain what was known of nickel steel suitable for 
bridge building which had not been recorded in print, the writer, in 
company with Mr. Osborn, made trips to Pittsburg and Pencoyd to 
consult with various officers and employees of the Carnegie Steel Com- 
pany and the American Bridge Company. The result of these visits 
was very satisfactory; for much valuable information was thus ob- 
tained, and some material from two old melts of nickel steel was se- 
cured for the preliminary testing. 

At the initiation of the work, a lay-out of tests covering the entire 
proposed investigation was prepared, and this was followed quite 
closely; nevertheless, as the testing proceeded, it appeared advisable 
to drop certain tests and to add others. One large group had to be 
abandoned on account of expense, as it involved the provision of an- 
other large melt of nickel steel; but from a number of tests on some 
other nickel steels the necessary deductions could be made. 

The initial lay-out of the work involved the following: 

1st. — The determination of the proportions of nickel and carbon 
that will provide a steel of the highest ultimate strength and elastic 
limit, consistent with perfectly satisfactory manipulation of the metal 
in the shops, and with absolute safety in the operation of structures 
built of it; 

2d. — The making of a complete series of tests of the nickel steel 
or steels adopted as standard, and a corresponding series of tests of 
the ordinary carbon steel used in the manufacture of bridges, so as to 
determine the values of all intensities of working stresses for nickel 
steel in bridges ; 

Sd. — The preparation of a thorough and minutely detailed set of 
specifications governing the designing of railroad and highway bridges 
in nickel steel; 

Jf.th. — The preparation of diagrams of weights of metal per linear 
foot of span, for all ordinary types of both single-track and double- 
track railway bridges, and for all spans, from that of the shortest plate 
girder up to that of the longest practicable cantilever, all designed 



NICKEL STEEL VOll rilflDGES 103 

for modern live loads and according to standard specifications, these 
diagrams to establish the weight curves for structures entirely of 
carbon steel, those entirely of nickel steel, and those of mixed nickel 
and carbon steels, in which the latter material is used for all parts 
where there would be no advantage in having the stronger metal; 

5th. — The preparation of a set of diagrams giving in comparison 
the total actual costs per linear foot of span for metal erected in 
nickel-steel and carbon-steel structures, and in mixed-steel and carbon- 
steel structures, for all the bridges of which the weights were dia- 
grammed, and for all possible pound prices for carbon steel erected 
(varying by ^ cent), and for all possible variations in pound prices of 
manufactured nickel steel and manufactured carbon steel delivered 
at the bridge site. These diagrams would enable anyone to perceive 
at a glance the economics of nickel steel for all standard types of rail- 
road bridges and for all possible conditions of the metal market; and, 
moreover, they would be good for all time, provided the characteristics 
of the two metals remained constant. 

The initial step in making the investigation was, of course, to de- 
termine the best percentage of nickel to use. At first thought, this 
appeared to be purely a commercial question, the idea being that the 
greater the percentage of nickel the greater the strength of the alloy, 
but, at the same time, the greater the cost per pound, and that, con- 
sequently, there must be some one percentage of nickel which is more 
economical than any other. This is true enough, but the enquiries 
made at Pittsburg indicated that when the percentage of nickel exceeds 
a certain amount (then not determined definitely) the material be- 
comes too refractory for shop manipulation or too expensive in manu- 
facturing. 

At the outset the writer had expected that it would be necessary 
to make a number of small special melts of nickel steel containing 
varying proportions of nickel, carbon, and possibly manganese, in 
order to settle the question of the best composition for the alloy; but 
the enquiries referred to and the preliminary tests on the specimens 
of nickel steel secured enabled him to determine this point without 
that necessity. Curiously enough, the superior limit of nickel for 
workability in bridge shops varies but little, if any, from the economic 
amount based on present prices of steel and nickel. For plates and 
shapes, which form the principal part of bridge superstructures, any 
materially greater percentage of nickel than 3J renders the alloy too 



104 NICKEL STEEL FOR BRIDGES 

refractory for the various shop manipulations to which it would be 
subjected in manufacturing it into bridges, although superstructures 
built of steel still higher in nickel would be perfectly safe and satis- 
factory for operation. 

The writer learned also, both by enquiries and by the preliminary 
tests, the best percentages of carbon to adopt for nickel steels of the 
various kinds required in bridge building; and he is of the opinion 
that any elaborate investigation of this question that might be made 
would not modify his conclusions materially. These best percentages 
of carbon will be indicated later, in connection with the discussion of 
the tests. For the present it will suffice to make the general statement 
that the greater the percentage of nickel in the alloy the greater the 
permissible amount of carbon, as far as safety against brittleness is 
concerned; and that, as the addition of carbon to the steel costs noth- 
ing and adds materially to both the ultimate strength and the elastic 
limit of the metal, it is economical to use as much of it as a proper 
consideration of workability in the shop and safety in operation of 
structure will permit. 

After making the enquiries at Pittsburg and Pencoyd, the writer, 
in consultation with Mr. Osborn, adopted the following maximum per- 
centages of impurities in the nickel steel recommended for bridge work : 

Phosphorus 0.03% 

Sulphur 0.04% 

Silicon 0.04% 

At the same time it was decided to leave acid open-hearth steel out 
of consideration, and to assume that structural nickel steel will be 
manufactured exclusively by the basic open-hearth process. 

As for the proper percentage of manganese, it was decided at first 
that this should be limited to about 0.6% ; but, later, it was found 
advisable to increase this to 0.75% for plate-and-shape steel, and to 
0.85% for eye-bar steel. Manganese adds to the strength of the alloy 
and costs comparatively little, but its use in excess renders the metal 
unduly hard. 

The following is a list of the comparative tests of nickel steel and 
carbon steel which were made in the investigation: 

A. — Specimen tests for tension, showing the elastic limit, ulti- 
mate strength, elongation, and reduction of area for the 
several kinds of steel required. 



NICKEL STEEL FOR BRIDGES 105 

B. — Bending tests on specimens, both plain and nicked, to de- 
termine the angle of bend and the character of fracture, 

C. — Bending on pins, to ascertain the extreme fiber stress for 
both elastic limit and ultimate strength, 

D. — Bearing on pins and rivets, to determine the proper intensi- 
ties of working stresses for bearing, 

E. — Shear on rivets, to find the correct intensity of working 
stress for shearing, 

F. — Tests for torsion, to ascertain the safe limit of extreme 
fiber stress on shafts of operating machinery, 

G. — Impact tests, to learn the relative resiliences of nickel steel 
and carbon steel, 

//.- — Combined impact and tension tests, to show the comparative 
resistances of the two metals for this combination of con- 
ditions, 

I. — Tests of full-sized eye-bars, 

J. — Tests of full-sized columns, both short and long, to aid in 
the preparation of proper formulas for nickel-steel struts, 

K. — Hammering-flat tests, 

L. — Drifting tests, 

M. — Close-punching tests, 

N. — Planing tests, 

0. — Drilling tests, 

P. — Chipping tests, 

Q. — Filing tests, 

R. — Specific gravity tests, 

S. — Coefficient of elasticity tests, 

T. — Corrosion tests, to determine the comparative resistances 
of the two metals to the attacks of acid, damp ielt, 
smoke, and wet cinders. 

Analyses of the metal from the two old melts of nickel steel ob- 
tained for testing, and the medium-carbon steel used for comparison, 
gave the results shown in Table 1. 

The percentages in Table 1 were obtained from drillings; but 
analyses of samples taken from the ladle gave, for the low-nickel steel, 
manganese, 0.68, carbon, 0.35, and nickel, 3.20; and for the high- 
nickel steel, manganese, 0.65, carbon, 0.42, and nickel, 4.22. 



lOG NICKEL STEEL FOR BRIDGES 

TABLE 1. — Results of Analyses of Steels. 



Character of alloy. 


Percentages of Ingredients. 


Sulphur. 


Phosphorus. 


Manganese. Carbon. 


Nickel. 


Low-nickel steel 


0.015 
0.014- 
0.021 


0.011 


0.65 0.39 


! 3.21 


High-nickel steel 


0.009 0.67 ' 0.463 


4.25 


Medium-carbon steel 


0.011 


0.46 0.275 

. 1 





The following is a condensed record of all the preliminary tests, 
which are given in full in Appendix C* 

Tensile Tests. 

An average of eleven tensile tests of the low-nickel steel gave 61 300 
lb. per sq. in. for the elastic limit and 99 300 lb. per sq. in. for the ulti- 
mate strength. An average of fourteen tests of the high-nickel steel 
gave 71 000 lb. per sq. in. for the elastic limit and 114 000 lb. per sq. 
in. for the ultimate strength. 

Bending Tests. 

The bends were made on unannealed specimens, 2 in. wide and 10.5 
in. long, with planed edges. The warm bends were made at ordinary 
shop temperatures, ranging from 55 to 70°, and the cold bends after 
the specimens had been immersed for 2 hours in a mixture of salt and 
ice. Four warm and four cold bends were made on each thickness of 
plates and angles of each of the three kinds of steel. The effect of the 
low temperature was quite marked. It seemed to be greater in the 
high- than in the low-nickel steel; but the tests were too few in num- 
ber to warrant stating this conclusively. 

All the low-nickel steel at the ordinary shop temperature was bent 
180° about two diameters; the high-nickel steel was bent 120° about 
three diameters. The cracks developed much more suddenly in the 
high-nickel steel and in the cold bends than in the low-nickel steel and 
in the warm bends, and were greater in magnitude. The bends of the 
low-nickel steel specimens were better than those of the high-nickel 
steel, but were not as good as those of the carbon-steel specimens. 

Impact Tests. 

There were four different series of impact tests. The first was 
made on specimens 21 in. long, | in. wide and 2 in. deep, reduced at 

* This appendix forms a part of the original paper, which may be found in the Society's 
Ijibrary. It consists of an itemized report of all the pi'eliminary tests by Mr. Sainiders. 



NICKEL STEEL FOR BRIDGES 107 

mid-length by two notches, each i in. wide and i in. deep, leaving an 
effective depth of only 1 in. Five pieces of each kind of steel were 
broken, the specimens being reversed after each blow. 

The apparatus, Fig. 1, was a rather crude affair, and was operated 
with much labor. The weight was a 50-lb. box, hoisted by a cord run- 
ning over a small drum. It was raised and lowered by a worm gear 
turned by a small crank. Additional weights, in units of 50 lb., were 
added by placing castings on top of the box. The weight was re- 
leased by disengaging a catch. 

The test pieces were supported on rounds, 12 in. between centers, 
and the load was transmitted through a tongue rounded at the bottom 
so as to bring the effect of the blow midway between the supports. 

Contrary to what had been gathered from the perusal of a number 
of papers on nickel steel, it was found by these experiments that the 
resilience of the nickel-steel alloys was considerably less than that of 
the carbon steel. All three steels were submitted to the same impact 
tests, the weight at first being 100 lb. and the drop increasing gradually 
from 6 to 24 in. ; then, after twelve blows, the weight was increased to 
150 lb. Calling the resulting average resilience of the carbon steel 
100, that for low-nickel steel was 87 and that for high-nickel steel 
was 73. 

This result might have been predicted, for turning over the speci- 
mens after each blow caused the total amount of abuse given to the 
metal to be a measure of its toughness and not of its resilience. 

The second series of impact tests was similar to the first, except 
that the specimens were not reversed. The result gave the following 
comparison of resiliences: 

Carbon steel 100 

Low-nickel steel 117 

High-nickel steel 109 

Because of the contradictory results of the two series of impact 
tests, it was decided to make a third, in order to determine the relative 
resistances of the three steels to a sudden blow. 

To make the tests, the apparatus shown in Fig. 2 was constructed. 
The hammer was forged round, with an eye-bolt in the top, and 
weighed 205 lb. It was raised by a traveling overhead crane. The 
specimens were similar to those used in the first series of tests, except 
that their length was reduced to 18 in., and there were ten of each of 
the three kinds of steel. 



108 NICKEL STEEL FOR BRIDGES 

No piece was struck more than one blow, and the height of drop 
first selected was such as would positively break the specimen. With 
the succeeding diminishing heights, an eflfort was made to approximate 
to that which would just break the piece. 

The comparative resiliences found by this series of tests were: 

Carbon steel 100 

Low-nickel steel 78 

High-nickel steel 71 

Thinking that the unsatisfactory showing of nickel steel under 
impact, as compared with carbon steel, and the diametrically divergent 
results of these impact tests, as compared with the results recorded by 
other experimenters, might be due to the notchings of the specimens, 
a fourth series, similar to the third, was made, but using plain bars 
of a section corresponding to that of the notched portion of the speci- 
mens previously tested, or | in. in width and 1 in. in depth. It proved 
necessary to use a 500-lb. hammer, as the 200-lb. hammer used pre- 
viously was entirely insufficient to break the specimens or to bend them 
materially. As it was, the tests were imperfect and inconclusive, for 
the high-nickel steel specimens broke under a drop of about 9.5 ft., 
those of low-nickel steel bent through an angle of from 120 to 130° 
under a drop of 12 or 13 ft. (only one piece breaking, and the others 
showing very slight indications of fracture), and those of carbon-steel 
bent under a drop of about 13 ft. without showing any signs of frac- 
ture. As the limits of the apparatus were reached, no attempt was 
made to carry the experiments any further. 

Although the result was imperfect, the test is quite satisfactory in 
that it shows clearly the great amount of abuse that low-nickel steel 
will stand, and that, in this particular, it is quite suitable for bridge 
building; although, of course, it is not as tough and ductile as carbon 
steel. Had the metal been at all brittle, the specimens would have 
been shattered instantly by so large a weight falling from so great a 
height, instead of first bending through such large angles. 

Up to the present, the general consensus of opinion has been that 
nickel steel has a higher resilience than carbon steel, and it was a mat- 
ter of surprise to the writer that three of his four series of impact ex- 
periments indicated the contrary. On this account a thorough dis- 
cussion of this feature of the paper is desirable. It may be concluded 



NICKEL STEKL FOR BRIDGES 109 

that at present the comparative resiliences of nickel steel and carbon 
steel are still an open question; and it is hoped that the discussion will 
throw sufficient light upon the matter to settle finally the existing un- 
certainty. 

It must not be forgotten that, in comparing the resiliences of 
nickel and carbon steels, if the strengths of the two metals be equal, 
or if their carbon contents be equal, nickel steel will give better re- 
sults than carbon steel ; and it is only when comparing nickel steel 
high in carbon with the ordinary low-carbon steel used in bridgework 
that the showing is ever unfavorable to the alloyed metal. 

CoiMRiNKD Impact and Tension Tests. 

The writer made an attempt to ascertain the effect of dropping a 
weight upon an eye-bar under tension, but the test was a failure. The 
weight (200 lb.) and the drop (4 ft.) were both too small. Although 
the bar was finally broken, it was impracticable to determine whether 
the ultimate strength was affected by the impact, as the piece de- 
veloped a pretty high resistance for medium-carbon steel. 

To obtain any results of value, it would have been necessary to set 
up a small pile-driver over the bar and drop a much larger weight 
from a greater height and with far greater frequency. After this 
experience it was concluded that "the game was not worth the candle," 
consequently the intention of carrying this test any further was 
abandoned. 

Later, the writer retained W. K. Hatt, Assoc. M. Am. Soc. C. E., 
of Purdue University, to make, among other tests, some experiments 
on combined tension and impact, and combined bending and impact. 
His full report is given in Appendix E.* The following extracts from 
it will show briefly his findings: 

"The impact tension tests were made on the machine in the Purdue 
Laboratory used by the writer in previous tests, and described in the 
Proceedings of the American Society for Testing Materials, Vol. 4, 
1904. The hammer used weighed 810 lb. This was allowed to drop a 
distance, h, in such a way that the energy of the blow was absorbed 
by the test piece. If the piece broke under the first blow, a record, 
taken on a revolving drum, made it possible to compute the energy 
left in the hammer after rupture. In nearly every case the piece was 
ruptured by one blow. 

* On file in the Library of the Society, but not reproduced in this paper. 



110 



NICKEL STEEL FOR BRIDGES 



"The impact flexure tests were made upon an electrically-operated 
impact machine. The hammer used weighed 55 lb. The 3 600-lb. 
base was supported by a 6-ft. concrete foundation resting upon undis- 
turbed gravel. The bars were tested on a span of 18 in., with the 
machined faces on the top and bottom. The loads were applied by 
allowing the hammer to drop upon the center of the beam from vary- 
ing heights, without turning the bar over between blows. The corre- 
sponding deflections were indicated by a recording pencil on the slowly 
revolving drum. The routine of each test was as follows : 

"With the hammer just touching the bar, a zero line was marked 
on the drum. The hammer was then allowed to fall from a height of 
i in. on the bar, the deflection being marked on the slowly revolving 
drum. These operations were repeated, increasing the height of drop 
i in. each time, until the elastic limit was well passed. When the 
elastic limit was reached, the material showed signs of set. This was 
marked on the drum by the recording pencil by turning the drum 
slightly after the vibrations of the bar had ceased. A typical drum 
record is shown on Fig. 3. None of the specimens, except the nicked 
bars, were broken in flexure. 

"Two of the nicked specimens, * * * one of carbon steel and 
one of nickel steel, were tested in a different manner from that out- 
lined above. The hammer was allowed to fall from a height of 18 in., 
then from a height of 24 in,, and then from a height sufficient to 
cause rupture. The drum records for these tests were similar to those 
obtained in the impact tension tests." 

Professor Hatt's results of impact tests are given in Table 2, which 
forms a part of his report. 

TABLE 2. — Results of Impact Tests. 



Impact Tenaioii. 

Elone:ation 

Contraction 

Rupture-work 

Impact Flexure. 
Plain bar. 
Deflection at blow from 3 in . . . 

Deflection at elastic limit 

Height of drop at elastic limit. 

Iinpdcl Flexure. 
Nicked bar. 
Deflection at blow from 3 in. . . 

Deflection at elastic limit 

Deflection at rupture 

Height of drop at elastic limit 
Height of drop at rupture 



Carbon. 



No. 



Aver- 



31.5 
58.9 
1736 



Maxi- 
mum. 



33.0 
60.8 
1910 



0.18 in. 0.31 in. 
0.21 " 0.83 '• 
3.70 " 3.50 " 



0.20 in, 
0.21 " 
1.23 " 
3.30 " 
11.10 " 



Mini- 
mum. 



31.0 
57.0 
1540 



0.16 in, 
0.16 " 
3.00 " 



(!.22 in. 
0.25 " 
1.40 " 
3.37 " I 
12.75 " 



0.17 i 

o.ir 

0.9 

3.25 

9.50 



Nickel (3.5%). 



No. 



Aver- 
age. 



16.5 
49.7 
2 198 



Maxi- 
mum. 



! 54 
2300 



5 0.09 in. 0.11 in. 
5 0,26 " 0.30 " 
5 8.10 " 9.00 " 



0.12 in. 0.14 in, 
0.23 •• 0.24 " 
0.62 " 0.70 " 
6.10 " ; 6.25 '• 
15.60 " 117.00 •• 



Mini- 
mum. 



0.09 in. 
0.22 " 
7.25 " 



0.12 i 
0.21 
0..57 
6.00 
14.5 



NICKKL STEEL FOR BRIDGES 111 

Table 2 shows that, in the impact tension tests, while the elonga- 
tion of carbon steel was nearly twice as great as that of nickel steel 
and the contraction was more than 20% greater, nickel steel had a 
resilience 26% higher than carbon steel — a result almost entirely at 
variance with the writer's findings. 

Table 2 also shows, in the impact flexure on plain bars, an increase 
in resilience of nickel steel over carbon steel equal to 120% at the 
elastic limit, and in the case of nicked bars an increase of 85% at 
the elastic limit and 40% at the point of rupture. 

Hammering-Flat Tests. 

The hammering-flat tests were quite conclusive, and proved that 
both the nickel steels compare very favorably with carbon steel in their 
ability to permit of being hammered flat. 

Drifting Tests. 

The drifting tests, as anticipated, show that nickel steels will not 
stand quite as much drifting as medium-carbon steel, nevertheless, 
both grades would comply with the drifting requirements for high- 
carbon steel given in standard bridge specifications. 

Close-Punching Tests. 

The experiments show that both the nickel steels withstood close- 
punching quite as well as the carbon steel, in fact, in one respect bet- 
ter, because the flow of metal was markedly less in the stronger steels, 
and the holes were smoother and more regular. 

Shop-Tooling Tests. 

The shop-tooling tests included planing, drilling, chipping, and 
filing. They were quite satisfactory, and gave the best comparative 
results of all the tests made. They showed conclusively that high- 
nickel steel is unfit for bridge building (except for eye-bars, pins, and 
rollers), not because of any inherent defect in the alloy, but because 
of its severity on the tools. To manipulate such metal in the shops, 
all machinery and tools would have to be built upon a more substantial 
basis than for carbon-steel work, and a sti'onger metal for punches 
and drills would have to be adopted. 



112 NICKEL STEEL FOR BRIDGES 

As for low-nickel steel, the tests showed that this alloy is all right 
for shop manipulation, but the time reqviired to do a certain amount of 
work upon it is greater than that required to do the corresponding 
work upon carbon steel. 

Corrosion Tests. 

The corrosion tests included immersion in a weak solution of sul- 
phuric acid, in damp salt, in locomotive fumes, and in wet cinders. In 
order to avoid the possibility of accident, the writer decided to have 
these tests made in duplicate — in his own office, at Kansas City, and 
in the office of the Osborn Engineering Company, at Cleveland. 
Owing to failure to ensure in advance that the methods of testing 
would be exactly alike in detail, the results are not truly comparable. 

All the plates tested were originally 6 by 6 by ^ in., and weighed 
about 5 lb. each. Three plates (one for each kind of steel) were used 
for each test, or twelve in all for each set of experiments. 

Acid Test. — Mr. Osborn used a 1% solution of sulphuric acid, 
with a new bath every two weeks, and he did not clean the specimens; 
while the writer used a 2% solution, changed the bath every week, and 
scraped the specimens each time the bath was changed. The results 
of these tests are recorded on Fig. 4. From the writer's record it will 
be seen that in 160 days 94% of the high-nickel steel was dissolved, 
thus ending the experiment, and that in the same time the low-nickel 
steel lost 89% and the carbon steel 52 per cent. Mr. Oshorn found in 
890 days the following losses of weight: 13.2, 12.9, and 11.7%, re- 
spectively. 

Both investigators were surprised by the results of this test, be- 
cause they had been told by one of the Carnegie Steel Company's 
superintendents that the pickle test would show up very favorably for 
nickel steel, as he knew by having made the experiment. 

This test, however, proves nothing against the use of nickel steel 
for bridges, because those structures are never exposed to such a con- 
dition or to any condition even approximating to it in severity, and 
because the other three corrosion tests, which agree in character more 
closely with the actual exposures of bridge metal, indicate a decided 
superiority of nickel steel over carbon steel. 

Salt Tests. — As shown by Fig. 5, the writer's salt test extended 
over 522 days, at the end of which time the carbon steel had lost 



nicki:l steel for bridges 113 

2.65% of its weight, the low-nickel steel 0.72%, and the high-nickel 
steel 0.63%, showing that nickel steel resists the attacks of salt four 
times as well as carbon steel. Whether the two steels when immersed 
in sea water would offer the same comparative resistances to corrosion 
is hard to say, but it seems probable that they would. If so, this 
would be a strong point in favor of using nickel steel for ocean piers 
and for bridges located near salt water. 

Mr. Osborn's results for the salt test were : In 840 days the carbon 
steel hadf lost 3.3%, the low-nickel steel 3.1%, and the high-nickel 
steel 3.2 per cent. The curves at 880 days cross each other, but this 
is probably due to some error in the records, since for 800 days the 
three curves are substantially parallel. 

The great variation in the two sets of experiments must be due to 
the fact that the writer's plates were kept continuously in the solu- 
tion and Mr. Osborn's were not. While this difference in treatment 
might account for the variation in the total amount of corrosion for 
any one kind of plate, it does not explain why the comparative re- 
sistances of the three kinds of steel should differ so greatly in the two 
sets of experiments. 

Smoke Tests. — As shown by Fig. 6, the smoke test was continued 
by the writer for 250 days, when the plates were lost by the renewal 
of the floor system of the bridge beneath which they were suspended. 
His test at the end of that time showed a loss of 6.25% of its weight 
by the carbon steel, 4.5% by the low-nickel steel, and 4.4% by the 
high-nickel steel, indicating that nickel steel resists locomotive fumes 
about 40% better than carbon steel. 

Mr. Osborn found the following results: In 310 days the carbon 
steel lost 0.18%, the low-nickel steel 0.12%, and the high-nickel steel 
0.04 per cent. 

The great difference in the results of the two tests was certainly 
caused by Mr. Osborn's failure to clean the plates from time to time, 
for the coat of soot deposited on the metal tended to preserve it from 
further attacks of the fumes. 

Cinder Tests. — As shown by Fig. 7, the cinder test was continued 
by the writer for 520 days, at the end of which time the carbon steel 
had lost 3.45% of its weight, the low-nickel steel 2.5%, and the high- 
nickel steel 2.15 per cent. This is quite a favorable showing for the 
nickel steel. 



114 NICKEL STEEL FOR BRIDGES 

Mr. Osborn's cinder tests gave the following results : In 890 days ■ 
the carbon steel had lost 16.7%, the low-nickel steel 14.5%, and the 
high-nickel steel 14.1 per cent. Up to 740 days, however, the high- 
nickel steel corroded a little more than the low-nickel steel. The 
greater amounts of corrosion in the Osborn test must have been due 
to the nature of the cinders used. 

From the various corrosion tests made by both experimenters, it 
may be concluded that nickel steel resists decidedly better than carbon 
steel the attacks of all substances which ordinarily tend to injure the 
metal-work of bridges. 

EivET Steel. 

While the various preliminary tests were being made, a thorough 
investigation of the qualities of a suitable nickel steel for rivets was 
carried on, using specimens of old melts, in order to avoid the ex- 
pense of making a special melt or melts for the purpose. While none 
of the five or six melts tested proved exactly suitable for rivets, enough 
was learned by the experiments to determine closely the proper pro- 
portions of the various constituents of rivet nickel steel. All the speci- 
mens tested were good enough for the purpose, except in one important 
particular, namely, the rivets were too difficult to cut out, the reason 
being that the carbon contents were too high. 

In the writer's opinion, ideal rivet nickel steel should have about 
the following composition : 

Nickel 3.5 per cent. 

Carbon 0.12 to 0.18 per cent. 

Manganese 0.55 to 0.65 per cent. 

Phosphorus 0.03%, maximum. 

Sulphur 0.04%, maximum. 

Silicon 0.04%, maximum. 

It is possible that a small economy might be effected by lowering 
the percentage of nickel and increasing that of carbon, and in time 
this may be done; but at present it does not seem worth while to make 
the nickel content less than that in plate-and-slmpe steel. The writer 
is of the opinion that by experimenting it could probably be shown 
that a rivet steel containing from 2 to 2.5% of nickel and from 20 to 
25 points of carbon would answer every purpose; but this is only a 
surmise on his part. It is possible, though, that such a high percentage 



NICKEL STEEL FOR BRIDGES 115 

of carbon would cause the rivets to burn too easily, but the writer's 
experiments gave no such indication. 

In addition to the ordinary tests for tension and bending, a num- 
ber of special tests on rivet nickel steel were made, among others, the 
following : 

A. — Driving two rows of nickel-steel rivets and two of carbon- 
steel rivets so as to connect four thicknesses of i by 12-in. by 3 ft. 
6-in. plates, reaming one row of holes for each kind, and noting how 
the nickel-steel rivets drove, as compared with the carbon-steel rivets: 

The result was that no difference whatsoever could be noticed in 
the difficulty of driving or the excellence of the rivets, and this was 
found to be true throughout the entire series of experiments on nickel 
steel and carbon steel. 

B. — Cutting off by hand the heads of most of the rivets mentioned 
in A, and driving them out of their holes, so as to obtain a proper 
comparison by noting the times required : 

The results of this test were very erratic, but, in general, it was 
found that the higher in carbon the nickel steel the longer it took to 
cut off the heads. When a pneumatic chipper was used, it sometimes 
required twice as much time for nickel steel as for carbon steel. 

In backing rivets out of reamed holes, no difference was noted in 
the resistances of the two kinds of rivets; but, in punched holes, the 
nickel-steel rivets came out in three-quarters of the time required for 
backing out the carbon-steel rivets. This result is due to the fact 
that the punched holes in nickel steel are much smoother than those 
in carbon steel. In this test (and in others also) it was learned that 
nickel-steel rivets should be used in nickel-steel plates and carbon- 
steel rivets in carbon-steel plates; for, when cutting out nickel-steel 
rivets from carbon-steel plates, the material around the rivet holes 
was injured because of the greater hardness of the rivet metal. 

C. — Punching four i by 6 by 15-in. plates with two rows of badly- 
matched holes (but making the matching in the two rows correspond), 
assembling the plates, reaming the holes, and filling one row with 
nickel-steel rivets and the other with carbon-steel rivets, then planing 
off the sides of the plates up to the medial planes of the rivets, so as 
to show how well the holes were filled : 

It resulted that there was but little difference in the flow in the 
two kinds of steel, what little there was being in favor of carbon steel. 



116 NICKEL STEEL FOR BRIDGES 

D. — Flattening the ends of i-in. rivets, both cold and hot, and 
noting how the metal stands the abuse: 

The cold nickel-steel rivets flattened to xs in. and then split, while 
the ends of the carbon-steel rivets flattened to fV in. before splitting. 
Undoubtedly, there would have been a better showing with nickel steel 
of ideal composition for rivets. The hot ends of both steels flattened 
out very thin without cracking. 

E. — Flattening rivets, both cold and hot, end on : 

In this test the cold carbon-steel rivets compressed 67% and the 
cold nickel-steel rivets 45 per cent. The latter required a much greater 
number of blows and received more damage. In the case of the hot 
rivets, both specimens flattened to a thickness of ^ in. and to a diam- 
eter of 3i in. without showing signs of cracking. 

F. — Testing rivets in double shear: 

The result of the tests of rivets in double shear showed that the 
nickel steel was about 40% stronger than the carbon steel, the ulti- 
mate strength of rivet nickel steel in shear being about 59 000 lb. per 
sq. in., or about 75% of the tensile strength. 

G. — Testing riveted joints: 

The result of the test on riveted joints was to obtain an ultimate 
shear of 68 000 lb. per sq. in. on the nickel-steel rivets and 46 500 lb. 
per sq. in. on the carbon-steel rivets. The ratio of these two figures 
is about 1.46, which is a little greater than the ratio found for the 
shears on single rivets. 

All things considered, the writer favors adopting rivet diameters 
for nickel-steelwork \ in. greater than those customary for carbon- 
steelwork. There are three good reasons for this, namely: 

First. — The increase of strength of nickel-steel rivets over carbon- 
steel rivets is not commensurate with the increase of strength in the 
plate-and-shape metal. This necessitates either a greater propor- 
tionate number of rivets or a greater rivet diameter in nickel-steel- 
work. 

Second. — The larger rivets keep hotter, and consequently fill the 
holes better. 

Third. — The strength of the punches is increased about one-third, 
while the work of punching each hole is increased only about 13 per 
cent. This reduces the danger from broken punches, and makes the 
work in punching nickel steel probably not more hazardous than that 
in punching carbon steel. 



NICKEL STEEL FOR BRIDGES 



117 



Eye-Bars of High-Nickel Steel. 

In order to determine approximately the kind of material for eye- 
bars the high-nickel steel would make, there were cut from the |-in. 
plates two strips, 5 in. wide, and these were forged into eye-bars, 
annealed, and tested. 

The elastic limit of the first bar broken was lost. Each bar broke 
near the neck, at a point where the temperature stresses are the 
greatest, and this, in connection with the low percentage of elongation 
in the body of the bar, would seem to indicate insufficient annealing. 
The breaking of both bars in the same place, however, may have been 
simply a coincidence. The fractures were 100% silky, and the re- 
ductions of area were excellent. 

The results of the two tests were as follows: 



Number of 

test. 


Elastic limit. 


Ultimate 
intensity. 


Percentage of 

elongation 

in 10 ft. 


Reduction 

of 

area. 


Fracture. 


Bar 1 


Lost 


105 900 
102 300 


7.4 
6.8 


46.90/0 
45.8"/o 


Silky, i cup. 
Silky, i cup. 


Bar 2 


Uncertain 





The unannealed specimen tests on the same plate gave an elastic 
limit of 71 100, an ultimate strength of 112 900, an elongation in 8 in. 
of 17.7%, and a reduction of area of 44.4 per cent. 

These tests are valuable in that they prove that satisfactory eye- 
bars can be manufactured from steel containing 4|% of nickel and 
about 45 points of carbon. 

Speci.\l Melts of Nickel Steel. 

Two special 40-ton melts of plate-and-shape steel were made for the 
writer by the Carnegie Steel Company, under the supervision of 
Albert Ladd Colby, M. Am. Soc. C. E., to whose care and attention is 
due the fact that the metal obtained was greatly superior in uni- 
formity to the low-nickel steel first tested. The composition of these 
special melts was determined by a number of independent chemists, 
with rather variable results, the general averages being as follows: 

Nickel 3.65%, carbon 0.39%, manganese 0.77%, sulphur 0.025%, 
phosphorus 0.01%, and silicon 0.05 per cent. 

The metal from these melts was rolled into plates and angles, 
some of which were used by the writer for testing. The remainder 



118 NICKEL STEEL FOR BRIDGES 

was distributed among the principal bridge shops of the country to 
undergo various manipulations, in order to let the manufacturers 
judge for themselves of the ease or difficulty of manufacturing it into 
bridges. This distribution was done liberally, even lavishly, and it is 
hoped that those who received the metal have submitted it to many 
practical shop tests, and that they will give the results thereof to the 
engineering profession in the discussion of this paper. 

The following is a list of the companies to whom material was 
thus sent: 

Dominion Bridge Co., Ltd., Montreal, Que.; 

Riter-Conley Mfg. Co., Pittsburg, Pa.; 

The Canadian Bridge Co., Walkerville, Ont. ; 

The Hamilton Bridge Works Co., Ltd., Hamilton, Ont.; 

Milliken Bros., New York City; 

The King Bridge Co., Cleveland, Ohio; 

McClintic-Marshall Construction Co., Pittsburg, Pa.; 

Pennsylvania Steel Co., Steelton, Pa.; 

Pore River Ship and Engine Co., Quincy, Mass.; 

Newport News Shipbuilding and Dry Dock Co., Newport News, 
Va.; 

William Cramp and Sons Ship and Engine Building Co., Phila- 
delphia, Pa. 

It was decided to postpone making the special melt of eye-bar steel 
until after the plate-and-shape steel had been tested. This was a wise 
precaution, because the testing of the latter might have caused some 
modification of the composition of the eye-bar steel; but the result, 
unfortunately, was disastrous, as the eye-bar steel was never manu- 
factured. This failure reduces somewhat the value of this research, 
as compared with what was outlined when the experiments were first 
contemplated. The eye-bars for testing were manufactured from the 
plate-and-shape steel; and, although the results were very satisfactory, 
the writer is convinced that much stronger eye-bars can be readily 
manufactured by using steel higher in nickel, carbon, and manganese, 
and that they will be suitable in every way for use in bridges. 

The special plate-and-shape nickel steel was tested very thoroughly 
and carefully, both in specimens and full-sized bridge members. Ex- 
cept in the case of eye-bars, for each test of nickel steel there was 



NICKEL STEEL FOR BRIDGES 119 

made a corresponding test of medium-carbon bridge steel. The results 
are recorded at length in Mr. Saunders' report, which is given in the 
original paper,* and is reproduced in a slightly condensed form in 
Appendix A of this paper. 

From it the writer has prepared the following summary and de- 
ductions : „ 

Tensile Tests. 

The average elastic limit is about 62 200 lb. per sq. in., and the 
minimum may safely be taken at 60 000 lb. for specimens cut from the 
edge ; the average ultimate strength is about 107 300 lb. per sq. in., 
and the minimum may safely be taken at 105 000 lb. for edge speci- 
mens. For specimens taken from the edges of plates, the ultimate 
strength is about 3 000 lb. per sq. in. higher, and the elastic limit 
about 2 000 lb. per sq. in. higher than for specimens taken from the 
interior. This ruling does not appear to apply to eye-bar flats, for 
their interior metal is just as strong as, if not stronger than, the metal 
near the edges. 

The elongation increases slightly with the thickness of the piece, 
and the reduction of area decreases as the speed of the testing ma- 
chine increases. 

The elastic ratio (that is, the ratio of elastic limit to ultimate 
strength) decreases as the thickness of the metal in plates or shapes 
increases; but the conditions affecting the testing of eye-bar metal 
were such as to prevent any reliable deduction being made on this 
point in relation to flats. This elastic ratio, in general, varies from 
0.55 to 0.60, but occasionally there are cases that pass slightly beyond 
these limits. 

Comparing the nickel steel and the carbon steel (which was of the 
kind specified as medium steel in "De Pontibus"), it was found that 
the average ratio of elastic limits was 1.94 and that of ultimate 
strengths was 1.79. 

The minimum elongation in 8 in. was about 15% for nickel steel 
and 27% for carbon steel. 

The minimum reduction of area was about 41% for nickel steel 
and about 55% for carbon steel. 

It was found that thick plates and angles gave lower elastic limits 
and ultimate strengths than thin ones. This was already known to be 

*0n file in the Society's Library. 



120 NICKEL STEEL FOR BRIDGES 

true for carbon steel, but the difference is so marked in nickel steel 
that any standard specification for the strength of that metal will have 
to take into account the thickness of the piece tested. 

Difficulty was found in determining the elastic limit, or, more 
strictly speaking, the yield point; consequently, Mr. Saunders was 
compelled to make some arbitrary assumptions, as shown by his report. 

It was noted during the tests that the nickel-steel angles were 
slightly inferior in strength to the nickel-steel plates, probably be- 
cause of the smaller amount of work the angles received in the rolling, 
and possibly because of the bend. It has been recognized for years in 
carbon steel that plate specimens give slightly better results than 
shape specimens. 

At the outset, the writer hoped to be able to obtain plate-and-shape 
steel which would have a minimum elastic limit of 64 000 lb., and that 
this might possibly be raised to 68 000 lb, by increasing the carbon 
content to 40 points. The experiments show an inferior limit of 60 000 
lb., which might indicate that Mr. Colby's two melts of plate-and-shape 
steel did not come up to the writer's anticipations; but such is not 
the case, for, when that metal was tested by the ordinary mill methods, 
the average elastic limit was 65 600 lb. per sq. in. As the manufac- 
turers' tests on the old melts were made by the usual mill methods, 
the writer based his judgment of possible results on that style of test- 
ing being adopted; but it was soon found that nickel steel should be 
tested much more slowly than has been customary for carbon steel, 
and the tests were conducted accordingly. In fact, the usual quick 
method of testing gives fictitiously great results for both the elastic 
limit and the ultimate strength of the latter metal. The elastic limit 
is especially affected. This feature of standard testing will be dis- 
cussed further, when the subject of eye-bars is treated. 

Tensile Tests on Punched, Reamed, and Punched-Riveted Specimens. 

The results obtained by this series of tensile tests are as follows: 
The punching alone raised the yield point in nickel steel from 2 

to 11%, and in carbon steel from 5 to 24%, at the same time lowering 

the ultimate strength of the nickel steel from 9 to 11% and that of 

the carbon steel from 6 to 7 per cent. 

This shows that the injurious effect of punching is a little more 

pronounced in nickel steel than in carbon steel. 



NICKEL STEEL FOR BRIDGES 121 

In the sub-punched-and-reamed specimens, both the nickel steel 
and the carbon steel regained all their ultimate strength, and in fact 
a little more. This shows the truth of what has been contended by the 
writer for a long time concerning reaming, namely, that it is the only 
method which insures truly first-class shopwork and the full strength 
of the manipulated metal. It also shows that, while this is true for 
carbon steel, it is especially so for nickel steel. 

The punched-riveted specimen tests indicate that driving the hot 
rivets into the punched holes increases still further the elastic limit 
and reduces the ultimate strength, thus furnishing another good rea- 
son for insisting upon sub-punching-and-reaming. The injurious 
effects of driving hot rivets in nickel steel and in carbon steel were 
about alike for the two metals. 

Bending Tests. 

The results of the bending tests on plain bars, when the bending 
is carefully and properly done,' indicate that the plate-and-shape steel 
may be bent 180° around a mandrel having a diameter equal to twice 
the thickness of the test piece without showing any signs of cracking 
the metal. 

Bending Tests on Punched, Reamed, and Punched-Riveted Speci- 
mens. 

A study of Fig. 2, Plate XV, of the inspector's report in Appendix 
A proves very clearly the injurious effects of punching on both 
nickel steel and carbon steel, for, while the reamed specimens of 
nickel steel bent on the average about 90°, and those of carbon steel 
180°, the punched specimens of nickel steel bent only about 50° and 
one of carbon steel only 84° ; and the punched-riveted specimens of 
nickel steel averaged only 38°, and one of the carbon-steel specimens 
bent only 62 degrees. 

This test is a strong confirmation of the necessity for sub-punching- 
and-reaming, and supports well the evidence of the tension tests on 
punched specimens. 

Drifting Tests. 

The result of the drifting tests proves that plate-and-shape nickel 
steel wil] withstand most of the ordinary drifting requirements for 
carbon-steel plates, but that, in the new specifications for nickel-steel 



122 "NICKEL STEEL FOR BRIDGES 

bridges, the required enlargement of the holes should be limited to 

40 per cent. 

Close-Punching Test. 

This test proves that nickel steel can be punched as closely as 
carbon steel, and, in fact, that it withstands punching better, in that 
the edges of the piece are bulged less and that the punchings are 
smooth and regular while in carbon steel they are rough and irregular. 
Fig. 1, Plate XVII, might lead one to think that this conclusion is not 
true, as one specimen shows several broken partitions. These were 
caused by the nervousness of the operator, who, fearing the breaking 
of the punch, dodged his head behind a protection at the instant of 
punching, and thus in several cases got the holes too close together. 
Nickel steel will stand any specification yet written for the close- 
punching of carbon steel. 

Hammering-Flat Test. 

The result of the hammering-flat test shows that nickel steel will 
stand hammering almost as well as carbon steel. 

Shop-Tooling Tests. 

The shop-tooling tests were quite elaborate, and covered every im- 
portant manipulation of metal in a bridge shop. It was found that 
nickel steel can be sawed with the usual saw and sheared with the 
usual shears. In punching, there was no special difficulty, but it is 
probable that more punches would have to be used for nickel steel than 
for carbon steel. If the diameters of all rivets in nickel steelwork are 
made J in. greater than those for carbon steelwork, as previously men- 
tioned, there should be no greater number of punches broken, because, 
as before stated, the strength of the punch, which varies as the square 
of its diameter, would be increased about one-third, while the work 
of punching would be increased only about 13 per cent. Moreover, it 
is practicable to improve materially the quality of the metal in the 
punches. 

In reaming, it was found that by the new process, termed "dry 
reaming," nickel steel can be reamed just as rapidly as carbon steel, 
but that the tools wear oiit more frequently. The old-fashioned method 
of reaming, in which oil or soapsuds are used for the lubricant, is not 
well adapted to nickel steel. 



NICKEL STEEL FOR BRIDGES 123 

In drilling, the same remarks apply as in the case of reaming. The 
ordinary drills fail quickly in nickel steel, but "blue-chip" drills stand 
the work very well. The writer's drilling tests were made without a 
lubricant. Had one been used, it is likely that better results would 
have been obtained. 

In planing nickel steel, it is not practicable to make cuts thicker 
than y,v in. ; while, in carbon steel, cuts of twice that thickness are 
usual. The effect on the tools is much greater for nickel steel than 
for carbon steel, and the amount of work that can be done on the 
former in a given time is much less than can be done on the latter — 
probably about one-half. 

In pneumatic chipping, it takes about 50% more time to make a 
certain length of cut in nickel steel than in carbon steel, and the 
amount of metal removed is slightly less for nickel steel. 

In hand chipping, a workman will cut in a given time only seven- 
tenths the length in nickel steel that he will in carbon steel, the thick- 
ness of the chips being the same. 

Bearing-on-Pins Test. 

It was found by the bearing-on-pins test that the plate-and-shape 
nickel steel is about 66% stronger than the carbon steel. In the case 
of bearing on rivets the excess of strength amounted to 83 per cent. 

Bending-on-Pins Test. 

Owing to the difficulty in testing round sections, square ones were 
substituted, and it was assumed that the comparative resistances for 
the two metals would be the same for both sections. It was found 
that nickel steel is about 85% stronger than carbon, steel in resistance 
to bending. Compression Tests of Struts. 

The compression tests of struts were most interesting and im- 
portant, especially as some engineers of high standing had surmised 
that there would be no economy in using nickel steel for long columns, 
because they anticipated that the two kinds of struts would deflect 
about the same for like total loads. It was found, however, that for 
short struts the nickel steel was 75% stronger than the carbon steel, 
and for struts of medium length 47% stronger. While these excesses 
of strength are not as great as those for tension, they are quite satis- 
factory. 



124 



NICKEL STEEL FOR BRIDGES 



Coefficient of Elasticity Test. 

This portion of the investigation was done with extreme care, 
three tests being made for each kind of steel, in which the average 
values of E proved to be 30 312 000 for nickel steel and 29 420 000 for 
carbon steel. For convenience in calculating, it will be advisable to 
take the coefficient of elasticity for nickel steel at 30 000 000. 

Tests of Eye-Bars and Eye-Bar Material. 

Unfortunately, all the eye-bars tested had to be rolled of plate- 
and-shape steel. Bars 1 and 2 in. thick were used, but no comparison 
was made with carbon-steel eye-bars of the same sizes. Eour 6-in., 
three 8-in., and one 16-in. bar were tested, and enough specimen tests 
of the metal were made to afford a proper average. The results of the 
tests are given in Table 3. 

TABLE 3. — Kesults of Tests of Eye-Bars and Eye-Bar Material. 

Six-Inch Bars. 



Specimens. 



Bars. 



Elastic limit. 



Ultimate . 



Unannealed 
I Annealed. . . 
I Unannealed 
I Annealed. . . 

I Unannealed 

I Annealed. . . 

I Unannealed 

Annealed.. . 



( Average 60 600 

(Minimum 56 900 

(Average 54 800 

/ Minimum 53 700 

j Average 103 100 

I Minimum 101 .300 

j Average 100 000 

1 Minimuni 98 700 

(Average 58 600 

/Minimum 54 700 

\ Average 52 500 

'( Minimum 50 700 

\ Average 98 900 

I Minimum 93 800 

( Average 95 200 

I Minimum 92 600 



\ Average 56 400 

/ Minimum. .. 55 400 

i Average.'.', .'"."l 01666 
"I Minimum.... 99 000 

I Average 51 700 

/ Minimum ... 48 30U 

\ Avei'a'ge'.'!!! "90266 
) Minimum.... 88 900 

49 400 
96 900 



Eight-Inch Bars. 



Elastic limit. 



Ultimate. 



Sixteen-Inch Bars. 



Elastic- limit. 



Ultimate , 



I Unannealed. 

-! 

I Annealetl 

1 Unannealed 



Annealed. 



(Average 59100 

(Minimum 58 500 

1 Average 55 600 I 

(Minimum 55 200 f 

( Average 103 600 

/ Minimum 103 100 

1 Average 103 200 1 

■| Minimwni 102 700 { 



NICKEL STEEL FOR BRIDGES 125 

The following deductions have been made from the results given in 
Table 3 : 

The average effect of the annealing of specimens is to reduce their 
elastic limits about 9% and their ultimate strengths about 3 per cent. 

The average elastic limits are for : 

Unannealed specimens 59 400 lb. per sq. in. 

Annealed specimens 54 300 '' " " " 

Full-sized bars 52 500" " " " 

The average ultimate strengths are for: 

Unannealed specimens 101 900 lb. per sq. in. 

Annealed specimens 99 500 '' " " " 

FuU-sized bars 96 000" " " " 

The average loss in elastic limit between the unannealed speci- 
mens and full-sized bars is about 11%, and the corresponding loss in 
ultimate strength is about 6 per cent. It is more than likely that, in 
the future manufacture of nickel-steel eye-bars, these variations be- 
tween the elastic limit and the ultimate strength of specimens and of 
full-sized bars will be lessened materially by a careful study of the 
process of annealing. 

Cold-Pressed Threads Test. 

An unsuccessful test on bolts of rivet nickel steel with cold- 
pressed threads was made and repeated. It seemed to be impracticable 
to manufacture such a bolt so as to have it break in the body before 
failing in the threaded portion; consequently, it was decided that 
nickel steel is not a proper material for the fabrication of bolts with 
cold-pressed threads. It is possible, though, that a nickel steel with 
a very low carbon content would work satisfactorily. 

Tests for Specific Gravity. 

Some very careful tests were made of the specific gravities of the 
two nickel steels and carbon steel, and it was found that the low-nickel 
steel was 0.017% lighter than carbon steel, and the high-nickel steel 
0.043% lighter. As these variations are extremely small, it will be 
proper to assume the weight of nickel steel to be exactly the same as 
that of carbon steel, when estimating weights of metal in bridges. 



126 NICKEL STEEL FOR BRIDGES 



Tests on Torsion.* 



Professor W. K. Hatt made for the writer a number of experiments 
upon rods 1 in. in diameter and 36 in. long, the nickel steel being 
standard "plate-and-shape" steel, and the carbon steel that ordinarily 
used in bridgework. He found the modulus of elasticity in torsion to 
be 11 760 000 for carbon steel and 11 450 000 for iiickel steel, the corre- 
sponding values of the modulus of rupture being 60 000 and 93 000, 
showing that low-nickel steel is about 55% stronger than carbon steel 
in its resistance to rupture, but practically equal to it in resisting 
distortion from twisting. Figs. 8 and 9 illustrate the results of Pro- 
fessor Hatt's torsion tests. 

Specifications for Nickel-Steel Bridges. 

From the results of the preceding investigations it is practicable 
to write specifications for nickel-steel bridges which will possess the 
same strength, rigidity, and general excellence of design as the best 
carbon-steel bridges that are being built to-day. 

In preparing such specifications, the writer has adopted as a 
standard those given in "De Pontibus," modifying only those portions 
which are affected by the change of metal; and, in order to limit the 
length of this paper as much as possible, he presents here only the 
modified portions. These are all that are necessary, for the main 
object of the specifications is to obtain the weights of metal from 
which the principal diagrams of this paper were prepared. Should 
anyone desire to design nickel-steel bridges by the use of these speci- 
fications, he can do so by combining them properly with those of "De 
Pontibus." Moreover, if the result of this investigation is to bring 
nickel-steel bridges into favor, it is the writer's intention to prepare 
and make public complete specifications for designing such structures. 

Metal. — All rolled steel shall be made by the basic open-hearth 
process. 

In bridges composed entirely of nickel steel, the eye-bars, pins, 
and rollers shall be made of "eye-bar steel" ; the rivets, bolts and all 
adjustable members sliall be made of "rivet steel"; and all other por- 
tions, except castings, shall be made of "plate-and-shape steel." 

* For a complete record of these tests see Appendix E of the original pappr in the 
Society's Library. It contains Pi-ofessor Hatt's report. 



NICKEL STEEL FOR BRIDGES 



137 



In bridges of mixed nickel and carbon steels, the floor system, the 
lateral system, and those parts of trusses of minor importance, such 
as struts having a large excess of section above the theoretical re- 
quirements, lacing bars, and stay-plates, may be made of carbon steel; 
but the floor system shall, preferably, be of nickel steel. 

TABLE 4. — Composition of Eolled Steel. 





Percentages. 


Inerredients. 


Rivet steel. 


Plate-and-shape steel. 


Eye-bar steel. 


Nickel 


a.50 (3.25 to 3.75) 
0.15 (0.12 to 0.18) 
0.(13 maximum 
0.04 maximum 
0.04 maximum 
0.60 (0.55 to 0.65) 


3.50 (3.25 to 3.751 
0.38 (0.34 to 0.42) 
0.03 maxmium 
0.04 maximum 
0.04 maximum 
0.70 (0.65 to 0.75) 


4.25 (4.00 to 4.50) 


Carbon 


0.45 (0.40 to 0.50) 


Phosphorus 

Sulphur 

Silicon 

Manganese 


0.03 maximum 
0.04 maximum 
0.04 maximum 
0.80 (0.75 to 0.85) 



As the manufacturer will have to keep the elastic limit and the 
ultimate strength i;p to certain minima, he will be allowed some 
liberty in the amounts of carbon to use in order to produce the re- 
quired results, but he is not to attempt to obtain such results by in- 
creasing the manganese, nor will he, under any circumstances, be per- 
mitted to pass the limits of phosphorus, sulphur, or silicon ; in fact, 
these limits should be kept as much below the specified amounts as 
practicable, because these elements are all detrimental to the metal. 
Preferably, the amounts of nickel should be kept within the limits 
set; but, in case of necessity, the latter, with the written permission 
of the Engineer, may be varied from. 

Method of Determining Elastic Limit. — The elastic limit for speci- 
men tests shall be assumed to be the load on the specimen producing 
a permanent set of 0.01 in. in a gauge length of 8 in., the amount of 
set being measured by fine dividers with no load on the specimen, or 
being taken from an autographic record at the intersection with the 
stress-strain curve of a line drawn parallel to and 0.01 in. away from 
the straight portion of the record. 

The elastic limit for tests of full-sized eye-bars shall be assumed 
to be the load on the bar producing a permanent set of 0.04 in. in a 
gauge length of 20 ft. ; or a proportionate set for shorter lengths, the 
amount of set being measured by an extensometer of approved design, 
with no load on the eye-bar. 



128 NICKEL STEEL FOR BRIDGES 

Tensile Strength. — The ultimate tensile strength per square inch 
on unannealed test pieces for all three kinds of rolled nickel steel used 
in structural metalwork shall be as follows : 

Eivet steel 70 000 to 80 000 1b. 

Plate-and-shape steel 105 000 " 120 000 " 

Eye-bar steel 115 000 " 130 000 " 

The preceding figures are for test pieces taken from the edge of the 
piece. In case the test pieces are taken from the interior, these figures 
may be reduced by 3 000 lb. each. 

The preceding figures also apply for all plates up to I in. in thick- 
ness and for all shapes up to f in. in thickness. For each additional 
I in. in thickness the ultimate strength may be reduced 1 500 lb. per 
sq. in., down to an inferior limit of 95 000 lb., for the thickest eye-bar 
flats. 

Elastic Limits. — The least allowable elastic limits per square inch 
obtained from unannealed test pieces shall be as follows: 

Rivet steel 45 000 lb. 

Plate-and-shape steel 60 000 " 

Eye-bar steel 65 000 " 

The preceding figures are for test pieces taken from the edge of the 
piece. In case the test pieces are taken from the interior, the fignires 
may be reduced by 2 000 lb. each. 

The preceding figures also apply for all plates up to J in. in thick- 
ness and for all shapes up to f in. in thickness. For each additional 
I in. in thickness the elastic limit may be reduced 1 000 lb., down to 
a limit of 57 000 lb. per sq. in. for plate-and-shape and eye-bar steels. 

Elongation. — ^The percentages of elongation shall be obtained from 
the unannealed test pieces after breaking on an original length of 8 
in., in which length must occur the curve of reduction from stretch 
on both sides of the point of fracture. The least allowable elongations 
for the three kinds of rolled structural steel shall be as follows: 

Rivet steel 25 per cent. 

Plate-and-shape steel 15 " " 

Eye-bar steel 12 " " 

The preceding percentages apply to plates, shapes, and flats i in. 
thick or less. For thicker metal they are to be increased by unity for 
each increase of i in. in thickness. 



NICKEL STEEL FOR BRIDGES 129 

Bending Tests. — Specimens of rivet steel shall be capable of bend- 
ing, by either pressure or hammering, to 180° and closing down flat 
upon themselves without cracking, when either hot or cold. 

Specimens of plate-and-shapc steel, when either hot or cold, shall 
be capable of bending by pressure 180° around a mandrel having a 
diameter equal to twice the thickness of the test piece, without showing 
signs of cracking on the convex side of the bend. 

Specimens of eye-bar steel, when similarly treated, shall be capable 
of bending by pressure 90° around a mandrel having a diameter equal 
to three times the thickness of the test piece, without showing signs 
of cracking on the convex side of the bend. 

Drifling Tests.— Vunchod rivot holes in plate-and-shape steel, 
pitched two diameters from a sheared edge, must stand drifting until 
their diameters are 40% greater than those of the original holes, and 
must show no signs of cracking the metal. 

The total taper of the drift-pins used for the testing shall not 
exceed 1 in 12. 

Fracture. — All broken test pieces for all three classes of steel, and 
all broken eye-bars must show a silky fracture of uniform color. 

Full-Sized Eye-Bars. — Full-sized eye-bars must show an ultimate 

tensile strength per square inch for the various thicknesses of metal 

as follows: . mK nnn iu 

l-m 105 000 lb. 

n-m 100 000 " 

2-in 95 000 " 

2^in. or greater 90 000 " 

The elongation shall not be less than 10% in a gauged length of 
10 ft. ; and the elastic limit shall not be less than 55% of the ultimate 
strength of the bar. 

Cast Steel. — All steel castings shall be made of open-hearth steel 
of the same composition as that specified for eye-bar steel, except 
that acid steel having a maximum limit for phosphorus of 0.06% may 
be used. The ultimate tensile strength shall vary within the limits 
of 110 000 and 130 000 lb. per sq. in. ; the elastic limit shall not be less 
than 55% of the ultimate tensile strength; and the elongation of test 
specimens in 2 in. shall not be less than 20 per cent. 

Impact Allowance Load. — The impact allowance load is to be a 
percentage of the equivalent uniform live load found by the following 
formulas : 



130 NICKEL STEEL FOR BRIDGES 

P= for railroad bridges, 

L + 500 

and P=^^ — r—, — ^ fo^ highway bridges, 
iv + loO 

where P is the percentage and L is the length, in feet, of span or 
portion of span covered by the live load, when the member considered 
is subjected to its maximum stress. 

Intensities of WorJcing Stresses. — The following intensities of 
working stresses (that is, pounds per square inch of cross-section) are 
to be used for all cases, except where wind loads are combined with 
other loads, in which cases the intensities are to be increased 25% ; 
but the sections shall not be less than those required by the stresses 
from all loads except wind. 

Tension on eye-bars 30 000 lb. 

Tension on plates and shapes in bottom chords, main 

diagonals, and laterals 28 000 " 

Tension on net section of plate-girder flanges (assum- 
ing one-eighth of the area of the web to act as 
part of each flange), on extreme fibers of rolled 
I-beams, and on shapes in body of suspenders, 
hip verticals, and hanger plates (there being 
50% increase of net area for section through 

eyes) 24 000 " 

Bending on pins 50 000 " 

Bearing on pins (measured upon the projection of 

the semi-intrados on a diametral plane) 38 000 " 

Bearing on rivets (measured similarly) 30 000 " 

Shear on pins 25 000 " 

Shear on rivets 14 000 " 

Shear on webs of plate girders (gross section) 17 000 " 

For field rivets, the intensities for bearing and shear are to be re- 
duced 20 per cent. 

Compression on top chords 30 000 — 120 

Compression on inclined end posts 30 000 — 140 

Compression on all other struts with fixed ends . 27 000 — 120 
Compression on all other struts with one or two 

hinged ends 27 000 — 160 ' 

r 



NICKEL STEEL FOR BRIDGES 131 

where I is the unsupported length of the strut, in inches, and r is its 
least radius of gyration, in inches. 

Compression on end stiffeners of plate girders 22 000 lb. 

For forked ends, the intensity of working stress shall be deter- 
mined by the formula, 

p=^ 15 000 — 450 - 
t 

where p is the greatest allowable intensity of working stress (impact 
being considered) ; I is the unsupported length, in inches, measuring 
from the center of the pin-hole to the center of the first transverse 
line of rivets beyond the point at which the full section of the member 
begins; and t is the total thickness of one jaw, in inches. 

The greatest allowable pressure upon expansion rollers of fixed 
spans, when impact is considered, shall be determined by the equation, 

p = 1 000 d, 
where p is the permissible pressure, in pounds per linear inch of roller, 
and d is the diameter of the latter, in inches. The preceding formula 
is to be used for rollers of swing spans with the span at rest, but, for 
the span in motion, the formula to be used is 

p = 400 d, 
where d is the mean diameter of the roller, in inches. 

In order to anticipate criticism, the writer will now^ take each fea- 
ture of the preceding specifications separately and show how the 
figures it contains were determined. 

Composition of Rolled Steel. — The composition of the rivet steel 
was fixed by the numerous experiments made on various specimens of 
rivet nickel steel, most of which w<jre objectionable in some particular. 
The essential requirements for a good rivet nickel steel are the follow- 
ing: 

First. — That it shall flow readily, so as to fill the holes properly, 

and that it shall head without splitting; 
Second. — That it is sufficiently soft to permit of the heads be- 
ing cut off and the rivets being backed out without un- 
due trouble and expense; 
Third. — That it shall have sufficient strength as compared with 
rivet carbon steel to make its use instead of the latter a 
decided desideratum. 



132 NICKEL STEEL FOR BRIDGES 

The rather wide limits of nickel, carbon, and manganese, of the 
specification, will assuredly permit manufacturers to fill these re- 
quirements. 

The composition of the plate-and-shape steel is that specified for 
the two special melts manufactured for this investigation under the 
direct supervision of Mr. Colby; and, while it is not claimed that ab- 
solute perfection was attained by this first trial, it is not likely that 
any great improvement in the characteristics of the future plate-and- 
shape steel, as compared with those of these melts, will be effected. 
This steel is strong, tough, reliable, and capable of being manipulated 
in the shops without danger to the workmen, and without unduly great 
expense. 

The satisfactory character of the composition of the eye-bar steel 
is not so well assured, as it was determined mainly by anticipation, the 
idea being to obtain as strong and hard a nickel steel as can be worked 
into eye-bars without running into brittleness. 

The existence of a brittle zone in nickel steels has been lately 
claimed by Dr. H. C. H. Carpenter, Mr. E. A. Hadfield, and Mr. 
Percy Longmuir, who presented a paper on the subject to the Institu- 
tion of Mechanical Engineers of England, in November, 1905.* 

These gentlemen find that with from 40 to 50 points of carbon, and 
from 75 to 100 points of manganese, a steel containing 4i% of nickel 
has lost none of its toughness or ductility, as compared with steels 
having smaller percentages of nickel; but that a steel containing 5% 
of nickel has lost both to a serious extent. Between these two per- 
centages there is probably a well-defined point of demarcation, but 
exactly where it is, can only be determined by further experiments. 

These should be made with the least possible delay. 

* 
They find also that with 7 or 8% of nickel the brittleness reaches 

a maximum, at 12% the resilience begins to increase, and at 20% 
the steel reaches its normal toughness. This is a very curious prop- 
erty of the alloy; and the knowledge acquired by these scientists, if 
corroborated by further experiments, will prove of great value to all 
metallurgists who are interested in the manufacture of nickel steel 
for bridges. The writer has heard, though, from good authority, that 
the best American practice in the manufacture of nickel steel does not 
agree with the findings of these experimenters, and that steel contain- 

* A r6sum6 of their investigations was published in The Etuihx-rr (London), November 
24th, 1005. 



NICKEL STEEL FOR BRIDGES 133 

ing 10% of nickel gave excellent results. On account of this dis- 
agreement of authorities, it is hoped that the discussion of this paper 
will hring out such a mass of evidence both pro and con that the ex- 
istence or non-existence of the so-called brittle zone will be firmly 
established. 

The recorded properties of the 4i% nickel steel experimented upon 
by these Englishmen were as follows: 

Carbon 0.40 per cent. 

Manganese 0.82 " " 

Bending test (on an anvil) 180 degrees. 

Yield point in tension 65 500 lb. per sq. in. 

"Ultimate strength in tension 109 100 " " " " 

Elongation in 2 in 20 per cent. 

Reduction of area 33 " " 

Modulus of elasticity 29 900 000 lb. 

The shock test, made by dropping a 46.7-lb. hammer about 14 in., 
showed that the 4.1% nickel steel absorbed a greater amount of energy 
before breaking than did any of the other steels tested, the amounts 
of nickel therein varying from zero to 20 per cent. This shows that, as 
far as impact is concerned, there is no objection whatsoever to 4^% 
nickel steel for eye-bars. 

Comparing the properties of this English 4|% nickel steel with 
those called for in the preceding specifications for eye-bar steel, it is 
seen that they strike an average for both nickel and manganese, touch 
the inferior limit for carbon, exceed somewhat the minimum elastic 
limit, and fall 6 000 lb. below the inferior limit for ultimate strength. 
Had the percentage of carbon been 0.45 instead of 0.40, this English 
steel would undoubtedly have had sufficient ultimate strength to meet 
the requirements of the specifications. 

As for the fitness of 4^% nickel steel to be manufactured into eye- 
bars, this is proved by the fact that for this investigation two eye-bars 
were fabricated from |-in. plates of 4|% nickel steel, as stated pre- 
viously. These eye-bars contained 46 points of carbon and 67 points 
of manganese, as determined by the chemist of the Osborn Engineer- 
ing Company; and the specimen tests gave elastic limits varying from 
69 200 to 78 200 lb. (determined by the drop of the beam at medium 
testing speed), ultimate tensile strengths varying from 112 100 to 



134 NICKEL STEEL FOR BRIDGES 

122 400 lb., percentages of elongation varying from 11.5 to 18.Y5 and 
averaging 16.8, and reductions of area varying (with one exception, 
where the fracture was unsatisfactory) from 39.5 to 47.1 per cent. 

As previously stated, the two eye-bars made from this material 
gave ultimate strengths of 102 300 and 105 900 lb. per sq. in.; but, 
unfortunately, the elastic limit in both cases was missed. The elonga- 
tion in 10 ft. was about 7%, and the reduction of area about 46 per 
cent. 

It must be remembered that these were not hona fide eye-bars, for 
they were manufactured from universal mill plates with planed edges. 
Had the edges been rolled, instead of planed, it is almost certain that 
better results would have been found, especially in the elongation. 
The amount of manganese was less than the specifications call for. 
Had it been increased, so as to agree with them, the ultimate strength 
of the specimens would probably not have fallen below their require- 
ments. 

As it was, the metal of which these eye-bars were made complied 
with the specifications in respect to elastic limit, gave an average 
ultimate strength only 300 lb. per sq. in. below the lowest requirement, 
and failed only in one case out of fourteen to give the demanded 
elongation (and then only by one-half of 1 per cent.). 

Considering that this was "picked-up" steel, and in view of the 
great superiority of Mr. Colby's low-nickel steel over the "picked-up" 
specimens of similar nickel steel, as shown by all the tests, it is fair 
to conclude that there will be no trouble whatsoever in obtaining eye- 
bar steel that will easily meet the preceding specifications, and at the 
same time will permit of satisfactory heads being forged with facility. 

The matter of annealing nickel-steel eye-bars, in order to obtain the 
best possible results, is one worthy of a full and immediate investiga- 
tion. The writer had intended to settle this question by experiments, 
but his failure to obtain a special melt of eye-bar steel prevented. It 
was his intention to test two thicknesses of eye-bars, namely, 2-in., and 
2i-in., from 6 to 8 in. wide, and 6 ft. long, in groups of three, by 
annealing at slightly varying temperatures, in order to determine from 
the average elastic limits and ultimate strengths of each group what 
temperature will give the best result for each thickness of metal. 

Should this series of tests on nickel steel for bridges ever be con- 
tinued beyond the limits of this investigation, one of the first steps 



NICKEL STEEL FOR BRIDGES 135 

would be to manufacture a special melt of eye-bar steel, make the usual 
specimen tests of the metal, so as to see that it is up to the require- 
ments, prepare a full supply of short eye-bars of various thicknesses, 
and determine finally the best annealing temperature for each thick- 
ness. 

The data concerning tests of nickel steel for the eye-bars of the 
Blaekwell's Island Bridge, given in Appendix B, were obtained through 
the courtesy of the American Bridge Company. From them the 
writer endeavored to obtain, by averages, approximate rates of varia- 
tion of elastic limit and ultimate strength with increased thickness of 
metal; but the attempt was a failure, as the results showed much 
irregularity, and really indicated that these characteristics of eye-bars 
are substantially independent of the thickness. Such a conclusion, 
however, would not be warranted, unless the metal of which all the 
eye-bars were composed was of practically uniform composition, which 
it was not, by any means, except in so far as the percentage of nickel 
was concerned. 

In order to obtain, if possible, some general information of value 
from these Blaekwell's Island Bridge tests, the writer has prepared 
averages of the chemical compositions; of the elastic limits, ultimate 
strengths, elongations, and reductions of area, of both unannealed and 
annealed specimens; and of the thicknesses, elastic limits, ultimate 
strengths, elongations, and reductions of area of full-sized bars, but, 
in making the averages, he rejected the elongations and reductions of 
area of those bars that broke in the eye. The results of these computa- 
tions are as follows: 

Chemical Analyses. 

Carbon 0.39 

Phosphorus 0.012 

Sulphur 0.03 

Manganese 0.71 

Nickel 3.39 

Unannealed Specimens. 

Elastic limit 60 610 lb. per sq. in. 

Ultimate strength 106 385 " " '' " 

Elongation 17.8% in 8 in. 

Keduction of area 29.9% 



136 NICKEL STEEL FOR BRIDGES 

Annealed Specimens. 

Elastic limit 54 377 lb. per sq. in. 

Ultimate strength 97 084 " " " " 

Elongation 22.5% in 8 in. 

Reduction of area 42.9% 

Full-Sized Eye-Bars. 

Thickness 1 f| in. 

Elastic limit 50 006 lb. per sq. in. 

Ultimate strength 88 818 " " " " 

Elongation 12.2% in 18 ft. 

Reduction of area 36.3% 

Comparing these average results with the corresponding figures 
given in the preceding "Specifications for Nickel-Steel Bridges," the 
following conclusions are reached: 

The impurities of phosphorus and sulphur are well within the 
limits of the specifications, the percentage of carbon is 6 points less, 
the percentage of manganese 9 points less, and the percentage of nickel 
86 points less; but, if this eye-bar metal be compared with the plate- 
and-shape steel of the specifications, the percentages of carbon and 
manganese are found to be almost the same, and the percentage of 
nickel only 11 points less; consequently, the average metal of the 
Blackwell's Island Bridge eye-bars is almost identical in composition 
with the plate-and-shape steel of the specifications, and with the special 
melts made for the writer's experiments. 

The average elastic limit and ultimate strength of unannealed 
specimens are very near the lower limits of the specifications for plate- 
and-shape steel, and considerably below those for eye-bar steel. Had 
the specifications for plate-and-shape steel been used for the manu- 
facture of these eye-bars, and had advantage been taken of the clauses 
permitting the lowering of the elastic limit and ultimate strength with 
the increased thickness of metal, there would have been no special 
difficulty in complying with the specifications, as far as the testing 
of unannealed specimens is concerned. 

Had the specifications for eye-bar metal been used for the manu- 
facture of these eye-bars, there would have been numerous rejections 
because of elastic limit, but very few, if any, on account of ultimate 
strength, as far as the testing of unannealed specimens is concerned. 



NICKKL S'l'KRL FOR BRIDGES 137 

As for elongation of unannealed specimens, it is evident that there 
would have been very little difficulty in complying with the specifica- 
tions for plate-and-shape steel, and none at all in complying with those 
for eye-bar steel. 

In respect to the tests of full-sized eye-bars, the average ultimate 
strength fell some 6 000 lb. below the requirements of the specifica- 
tions, and the average elastic limit about 2 500 lb. below. Of course, 
in individual instances, the discrepancies were much greater. 

Had the eye-bar specifications been drawn on the assumption that 
plate-and-shape steel was to be used for eye-bars, the limiting ulti- 
mate strength for 2-in. bars would have been about 87 500 lb. per sq. 
in., and the corresponding elastic limit about 48 000 lb. per sq. in. In 
every test the elastic limit would have been complied with, but in 35% 
of the tests the ultimate strength would have fallen short, sometimes 
materially so, owing to the fact that many bars broke in the eye. 

As for the elongation, no difficulty would have been found in com- 
plying with the specifications. 

Comparing the differences in elastic limit and ultimate strength 
between unannealed and annealed specimens, the averages of all the 
tests show, respectively, 6 233 and 9 301 lb., the corresponding per- 
centages being 10.3 and 8.7. The writer found in his experiments ap- 
proximately 9 and 3, which is a fairly close agreement. 

Concerning the diflferences in elastic limit and ultimate strength 
between unannealed specimens and full-sized eye-bars, the averages 
of all the tests show, respectively, 10 604 and 17 567 lb., the corre- 
sponding percentages being 17.5 and 16.4. The writer found in his 
experiments approximately 11 and 6, These discrepancies are mainly 
due to the fact that the inspector made his specimen test slowly, but 
possibly in part by more careful annealing of the eye-bars. No data 
were obtained from the American Bridge Company concerning the 
annealing of their nickel-steel eye-bars, but the writer hopes that some 
of its officers, in discussing this paper, will treat this point thoroughly. 

Method of Determining Elastic Limit.— The method described in 
the specifications for determining the elastic limit will suffice while 
the building of nickel-steel bridges is still in its infancy, because at 
first there will be ample time for testing the new metal; but, after- 
ward, when the manufacture of such structures from nickel steel be- 
comes general, the testing will have to be done on a more commercial 



138 NICKEL STEEL FOR BRIDGES 

basis, and the elastic limit will then have to be determined by the drop 
of the beam. 

For such conditions, this specification is suggested: 

"The elastic limit, in testing specimens, may be determined by the 
drop of the beam, according to the accepted practice in all steel mills, 
provided the speed of the machine or the movement of the head is not 
more than 1 in. in 3 min., up to the elastic limit, and that the weight 
is moved out by hand at a uniform speed sufficient to keep the beam 
in very light contact with the upper cross-piece. The speed, after the 
elastic limit is passed, shall not be greater than 1 in. in 30 sec. nor less 
than 1 in. in 1 min." 

The true elastic limit may be determined as just described, pro- 
vided the operator maintains a uniform movement of weight, as speci- 
fied. 

Little dependence can be placed on the elastic limits from specimen 
tests as usually made, because the speed of the machine is so rapid 
that the lever is kept pressed hard against the upper cross-bar, and 
thus the elastic limit is passed by 3 000 or 4 000 lb. per sq. in. before 
the beam has a chance to drop. The ultimate strength is exaggerated 
in the same manner, but not usually to the same extent as the elastic 
limit. In order to obtain proper and reliable records for all specimen 
tests, the limits of the speed, both before and after the elastic limit is 
passed, should be materially reduced, but not to such an extent as 
either to involve a hardship for the manufacturer or to cause the in- 
spectors excessive labor or trouble. 

Tensile Strength. — It has been shown previously, for this item, that 
there will be no special difficulty in filling the requirements of the 
specifications for the three kinds of nickel steel; especially if it be re- 
membered that the figures given as limits may be reduced 3 000 lb. 
each for specimens taken from the interior, and that there is a sliding- 
scale reduction for specimens exceeding a certain thickness. With 
these limitations, manufacturers should have no serious difficulty in 
obtaining in all cases the ultimate strength called for. 

Elastic Limits. — The preceding remarks upon the specification for 
ultimate strength apply to the elastic limit specification as well. 

Elongation. — The specifications governing elongation are by no 
means difficult to fill; indeed, they are possibly not severe enough, and 
in the future it may be advisable to raise the percentages somewhat. 



NICKEL STEEL FOR BRIDGES 139 

Bending Tests. — None of the requirements specified for bending are 
difficult to satisfy, provided the operator does not attack the metal 
brutally, but gives it a chance to show what it is worth by taking 
plenty of time and by seeing that the bend is not made too sharp. 

Drifting Tests. — Nor are the drifting-test requirements difficult 
to satisfy, if the operator will not strike too hard blows and if he will 
turn the plate over quite often. The enlargement of the holes should 
always be made gradually and with care. 

Fracture. — As practically all the test pieces in this series of tests 
complied with this requirement, there should be no difficulty in the 
future in living up to it. 

Full-Sized Eye-Bars. — As it is not likely that nickel-steel eye-bars 
will be made less than IJ in. or more than 2^ in. thick, it is not prob- 
able that there will be any special difficulty in obtaining ultimate 
strengths of 100 000 lb. for the smaller thickness, and 90 000 lb. for 
the greater thickness. The writer found, for the weakest of his 2-in. 
bars, an ultimate strength of 89 000 lb. per sq. in., and, as the metal 
specified for eye-bar steel is 10 000 lb. stronger in specimens than that 
of which his eye-bars were fabricated, there ought to be no difficulty 
in meeting the requirements. 

The elastic limit in nickel steel never falls below 55% of the ulti- 
mate strength. 

Cast Steel. — The specifications for cast steel are not difficult to 
fill, and if they should prove so in any particular it would be easy 
enough and perfectly legitimate to amend them, as castings form but 
a small portion of bridge material. 

Impact Allowance Loads. — The impact allowance loads are the 
same as those given in "De Pontibus," and their reliability has never 
been questioned; for it is generally conceded that they are decidedly 
in excess of the real impact. It was the writer's intention, when the 
book was written, to make them cover, not only the actual impact, but 
also small, unavoidable secondary stresses and slight inequalities of 
stress distribution. The various intensities of working stresses adopted 
were properly adjusted to these impact allowances. 

Tension on Eye-Bars. — There are several ways of checking the cor- 
rectness of the 30 000 lb. per sq. in. for the working stress on eye-bars. 
One is to take the ratio of least allowed elastic limits for full-sized 
eye-bars of nickel steel and of carbon steel and multiply it by the 



140 NICKEL STEEL EOR BRIDGES 

intensity given in "De Pontibus" for carbon-steel eye-bars, namely, 
18 000 lb. The result would be : 

49 500 X 18 000 .^, „,„ ,, 

2iT000 = "^ '"^ '''■ 

Another way is to use the ratio of least allowed ultimate strengths 
of full-sized eye-bars. The result would then be: 

5b UOU 
Still another check is to use the ratio of least allowable elastic 
limits for specimen tests, which would give: 

65 OOP X 18 OOP ^ 3g ^^^ „^ 

3o 000 

Or, adopting the ratio of least allowable ultimate strengths of 
specimens, there results: 

oU 000 

Averaging these four results gives 32 100 lb., which is well above 
the 30 000 lb. specified. 

Tension on Built Members. — The intensity given in the new speci- 
fications is 28 000 lb., and that for carbon steel from "De Pontibus" is 
16 000 lb. Applying the ratio of least allowable elastic limits, gives : 

60 000 X 16 000 _ 

35 000 " ■ 

Or, taking the ratio of least ultimate strengths: 

"' °""n ?ln' ""^ = 28 OOP lb. 

bO 000 

It is true that the first check shows a deficiency of 600 lb., but 
this small amount is fully compensated for by the greater resistance 
of nickel steel to the abuse which all metal receives in the shops. 

Tension on Net Section of Flanges of Beams, etc. — The intensity 
of tension on net section specified for nickel steel is 24 000 lb., and 
that for carbon steel is 14 000 lb. Applying the ratio of least elastic 
limits gives: 

«'L""«X,'i«^' = 24.)»0lb. 
35 000 

Bending on Pins. — The ratio of bending resistances on pins for 
plate-and-shape nickel steel and carbon steel found by the tests was 



NICKEL STEEL FOR BRIDGES 141 

1.85. As eye-bar steel is about 8% stronger, this ratio should be in- 
creased to 1.85 X 1-08 = 2.0, The working intensity for bending on 
pins of carbon steel is 27 000 lb., consequently, that for pins of high- 
nickel steel would be 2 X 27 000 = 54 000 lb., while the specifications 
call for only 50 000 lb. If the ratio of 1.85 were used, the intensity 
would be almost exactly 50 000 lb. 

Bearing on Pins. — For bearing on pins, the writer found, for plate- 
and-shape steel compared with carbon steel, a ratio of 1.66. The in- 
tensity for carbon steel is 22 000 lb., consequently, thfe application of 
the ratio would give 22 000 X 1-66 = 36 500 lb. Owing to the superior 
stiffness of nickel steel, and because the intensity given in "De 
Pontibus" is rather low for bearing on pins as compared with the other 
intensities of the specifications, the intensity for bearing on nickel- 
steel pins was taken at 38 000 lb. It must be remembered that this is 
not calculated for the high steel of the pin, but for the lower steel of 
the bearing. 

Bearing on Rivets. — The experiments show that for rivets the ratio 
of bearing stresses is 1.83. The intensity for carbon steel is 20 000 lb. 
Applying the ratio gives 20 000 X 1-83 = 36 600 lb. Another method 
of calculating this is to take the 38 000 lb. found for bearing on pins 
and multiply it by the ratio of least elastic limits of rivet nickel steel 
and plate-and-shape nickel steel, giving 38 000 X 45 000 -f- 60 000 = 
28 500 lb. As a compromise between these two widely varying results, 
a value of 30 000 lb. was adopted. 

Shear on Pins. — As no experiments were made on shear on pins, 
the proper intensity can be taken by proportion from the established 
bending intensities for both nickel-steel and carbon-steel pins and 
from the given working shear of 15 000 lb. per sq. in. for carbon-steel 
pins, thus S = 15 000 X 50 000 ^ 27 000 =- 27 800 lb. The amount 
adopted in the specifications is only 25 000 lb. 

Shear on Rivets. — The writer found the comparing ratio for ulti- 
mate strengths of nickel-steel rivets and carbon-steel rivets in shear 
to be about 1.4. Applying this to the intensity of working stress for 
carbon-steel rivets, namely, 10 000 lb., gives 14 000 lb. for the in- 
tensity of shear on nickel-steel rivets, which is the figure adopted for 
the specifications. 

Shear on Wehs of Plate Girders. — The intensity of working stress 
for carbon steel is 10 000 lb. ; and, as the ratio of elastic limits for 



p = 18 000 — 70 



142 NICKEL STEEL FOR BRIDGES 

plate-and-sliape steel and carbon steel is about 1.7, the intensity for 
shear on web plates of nickel steel should be 17 000 lb., as given in the 
specifications. 

Compression Formulas. — As the column tests on nickel steel were 
limited to six, the data for preparing formulas were very meager, but 
the comparing ratios found by the writer for long and for short 
columns, namely, 1.47 and 1.75 (the ratios of length to least radius of 
gyration for the two cases being, respectively, 27 and 81), sufficed for 
the establishment of empirical formulas similar to those of "De 
Pontibus." 

The short struts tested correspond to top-chord panel lengths, for 
which the carbon-steel formula is: 

I 
r 

The formula assumed for nickel-steel top chords in the specifica- 
tions is: 

p = 30 000 — 120 ^ 

r 

Testing this for = 30, gives 
r 

p = 15 900 for carbon steel, 
and p = 2Q 400 for nickel steel. 
The ratio of these values is 1.65, instead of 1.75 as shown by the experi- 
ments. 

For =^ 50, which is the usual limit for top-chord sections of rail- 

r 

road bridges, 

p = 14 500 for carbon steel, 
and p = 24 000 for nickel steel. 
The ratio of these values is 1.65. By interpolation from the experi- 
ments, this would have been about 1.64. 

These examples show that the new top-chord formula errs on the 

side of safety for small values of i and is just right for the usual 

values in top chords. 

The formula for inclined end posts of nickel steel was obtained 
approximately from that found for the top chords by simply varying 

the coefficient of by the ratio of the corresponding coefficients in the 
r 

formulas for carbon-steel struts. 



NICKEL STEEL FOR BRIDGES 143 

For all other nickel-steel columns with fixed ends the formula as- 
sumed in the specifications was: 

p = 27 000 — 120 ^ • 

r 

while that for carbon-steel struts is: , 

p = 16 000 — 60 -. 
r 

The usual average value of for such struts is about 80, for which 

r 

the formulas give: 

p = 17 400 for nickel steel, 
and p = 11 200 for carbon steel. 
The ratio of these values is 1.55, which is somewhat more than the 
ratio given by the experiments for hinged ends; but it would probably 
be amply safe for fixed ends. 

The formula for all other nickel-steel struts with one or two hinged 
ends was derived from the formula for similar struts with fixed ends 

by modifying the coefficient of in about the same ratio as exists in 

/' 

the two carbon-steel column formulas, namely, 80 -^ 60 = 1.33. Thus, 
120 X 1-33 = 160; which is the constant adopted in the nickel-steel 
column formula for hinged ends. 

Testing this for = 80 gives: 
r 

p = 14 200 for nickel steel, 
and p = 9 600 for carbon steel. 
The ratio of these values is 1.48, which agrees almost exactly with 
the results of the experiments. 

As soon as nickel steel is actually used for building bridges, it will 
become necessary to make some additional experiments on columns 
with both fixed and hinged ends for various lengths so as either to es- 
tablish some new formulas or to verify the preceding ones. Meanwhile, 
it will be perfectly safe to use the latter in designing nickel-steel 
bridges. 

Compression on End Stiffeners of Plate Girders. — The intensity 

for carbon-steel stiffeners is 14 000 lb., and as for end stiffeners is 

r 

small, say not to exceed 40, the safe ratio to use will be 1.68, hence 
the intensity for nickel-steel end stiffeners could be 14 000 X 1-68 = 
23 500 lb., while the specifications call for only 22 000 lb. 



144 NICKEL STEEL FOR BRIDGES 

Formula for Forked Ends. — As in forked ends is likely to run 

r 

as high as 80, the safe ratio will be about 3.5. Applying this to the 
carbon-steel formula, namely, 

p =: 10 000 — 300 ^ 
t ' 

gives p = 15 000 — 450 ' , 

which is the formula adopted for forked ends of nickel steel. 

Formula for Expansion Rollers. — The carbon-steel formula for ex- 
pansion rollers is p = 600 d. Applying the general ratio of strength, 
namely, 1.7, gives for the constant 1.7 X 600 = 1 040 ; hence the 
formula for nickel-steel rollers at rest was made p = 1 000 d. 

For rollers in motion, this was changed to p = 400 d. 

The only specification previously written for nickel steel in bridges, 
so far as the writer knows, is that of E. S. Buck, M. Am. Soc. C. E., 
for the Manhattan Bridge, issued by the Department of Bridges of the 
City of New York. In it the following intensities are specified for 
nickel-steel members for a combination of dead load, temperature, and 
congested live load, or for a combination of dead load, regular live 
load, temperature, and wind. 

Pounds per square inch. 
Tension in stiffening trusses 40 000 

Compression in stiffening trusses 40 000 — 150 — 

r 

Shear on rivets in stiffening trusses (field) 20 000 

Bearing on rivets in " " " 35 000 

As the writer's proposed specifications are for a combination of 
dead load, live load, impact-allowance load, and wind load, his allow- 
ance for impact, for the purpose of comparison, may be allowed to off- 
set Mr. Buck's allowance for temperature. For this combination, the 
writer's intensities of working stresses would be as follows : 

Pounds per square inch. 

Tension in stiffening trusses (eye-bars) 37 500 

Tension in stiffening trusses (shapes) 35 000 

Compression in stiffening trusses (chords) 37 500 — 150 — 

r 
Compression in stiffening trusses (webs), fixed 

ends 33 750 — 150 i 

r 



NICKEL STEEL FOR BRIDGES 145 

Compression in stiffening trusses (webs), hinged 

ends 33 750 — 200 ^ 

T 

Shear on rivets in stiffening trusses (field) 14 000 

Bearing on rivets in stiffening trusses (field) ... 30 000 

Comparing the two sets of figures, it will be seen that Mr. Buck 
has in every case stressed his nickel steel higher than the writer has 
stressed his, notwithstanding the fact that the requirements for 
strength of metal in Mr. Buck's specifications are decidedly less than 
in those of the writer. It is evident, therefore, that, as compared with 
the dicta of the sole present authority on the use of nickel steel in 
bridgework, the specifications herein presented for the designing of 
nickel-steel bridges err assuredly on the side of safety. It is true that 
Mr. Buck's specifications are for a very long span, and in consequence 
his intensities of working stresses are permitted to run high by the 
best established engineering practice; but it must be remembered that 
the writer's intensities, by the impact allowance of his specifications, 
are adjusted properly for spans of all lengths. 

In making this comparison the writer is not endeavoring to criti- 
cise Mr. Buck's specifications, but is simply anticipating possible 
criticism of his own on the plea of overstraining the nickel steel. 

On Figs. 10 to 21, inclusive, are given the weights of all ordinary 
single-track and double-track railway bridges for all spans up lo 
1 800 ft. There are included four types of cantilever bridges, namely, 
A, B, C, and D, as shown on Fig. 22. Type A is the most usual, and 
is suitable where only one long span is necessary. Type B is for a 
bridge of very great length, where two long main spans are required. 
Type C is for the case where the total length is greater than for Type 
A, but short as compared with that for Type B. As regards economy 
of metal. Type C comes next to Type A. Type D is intermediate be- 
tween Types B and C. These four types cover all the possible lay- 
outs of spans for cantilever bridges, or at least all that are consistent 
with good engineering practice. 

Class B of the "De Pontibus" specifications was adopted as the 
live load for simple-span bridges; and, for cantilever structures. Class 
R was used for the stringers, Class S for the floor-beams and the 
primary truss members, and Class U for the main truss members. All 



I'i6 NICKEL STEEL FOR BRIDGES 

the cantilever structures were assumed to have double tracks, as such 
bridges are now rarely, if ever, built with single tracks. 

For the purpose of record, all lay-outs for cantilever bridges were 
assumed to have the following constant proportions between the lengths 
of their various spans: 

Calling I the length of the main span of Type A, § I will l)e the 
length of the suspended span and xV I that of each cantilever arm and 
of each anchor arm. For the anchor span, when there is one, the 
length will be f Z. These proportions are all shown correctly to scale 
on Fig. 22. The plotted weights of metal per linear foot of span for 
the cantilever bridges are the average weights for the entire length of 
structure. 

The weights of metal per linear foot of span for the carbon-steel 
bridges were computed by using the specifications of "De Pontibus," 
and those for the nickel-steel and mixed-steel bridges by the specifica- 
tions of this paper combined with those of "De Pontibus." These 
weights are as accurate as they can well be made, and much time was 
spent by the writer's office force in calculating them. At some future 
time, after bridge building in nickel steel has been inaugurated, tho 
writer will give to the Profession curves of weights of metal per 
linear foot in nickel-steel, mixed-steel, and carbon-steel bridges for all 
his standard live loads; but, for the present, those offered in this paper 
will have to suffice. 

From the weights of metal per linear foot, given on Figs. 10 to 21, 
inclusive, from various assumed pound prices of carbon-steel bridges 
erected, and from various assumed differences in pound prices of 
superstructure metal delivered at site, in nickel steel and in carbon 
stoel, the costs in dollars per linear foot of span (for bridges of all 
types in "all nickel steel," "mixed steel," and "carbon steel") were 
plotted, and are given in Figs. 23 to 72, inclusive.* 

The range of pound prices for carbon steel erected is from 2.5 to 
5.5 cents for plate-girder bridges, and from 3.0 or 3.5 cents to 6.0 cents 
for all other bridges. These ranges are likely to include, for many 
years, all the conditions of the market for carbon-steel bridges erected ; 

* Only 50 of 154 flia?:rams are reproduced in this paper. Tlie others may be seen in the 
Society's Library. These 50, ho«-ever, are iiil lliat will usually be neeiled in the im- 
mediate future when comparing the costs of carbon-steel bridges with those of bridges 
built of nickel steel in all parts where the ado]ition ot the alloy is economical, and of carb.m 
steel in all other parts. 



NICKEL STEEL FOR BRIDGES 147 

although a combination of general prosperity and a distant and diffi- 
cult location might cause the superior limits to be passed occasionally. 
Such a contingency, however, is too remote to warrant the preparation 
of more diagrams than those that accompany this paper, especially 
since, in such a case, it would require only a few minutes to make by 
proportion the necessary correction for the special proposed structure. 

The differences in pound prices of nickel steelwork and carbon 
steelwork delivered at site range from 0.6 cent to 2.0 cents. These 
ought to be sufficient, for when the difference becomes as low as 0.6 
cent, these tables will have served their purpose and become obsolete, 
because then practically all bridges will be built of nickel steel; and 
even to-day a variation of 2 cents would indicate an excessive over- 
charge on the part of the manufacturers. In the case of the Manhat- 
tan Bridge, the difference bid by the contractors between nickel-steel 
eye-bars and carbon-steel eye-bars erected was 1.5 cents per lb.; con- 
sequently, the difference for the steels delivered at site would have 
been somewhat less than that amount. It is true that the difference 
in cost of manufacture of entire bridges in nickel steel and in carbon 
steel is somewhat greater than the corresponding difference in the case 
of eye-bars alone, but the variation is certainly not so great as h cent 
per lb. If a greater difference than 2 cents per lb. between the values 
of nickel steelwork and carbon steelwork delivered at site should occur 
in any case, it would be an easy matter to plot a curve to meet the 
condition by proportionate extension on the special diagram that is 
to be used. 

The plotted pound prices for carbon-steel bridges erected vary by 
i cent per lb. In case it is desired to assume any price intermediate 
between those on the diagrams, the comparison between costs per foot 
of nickel-steel, mixed-steel, and carbon-steel bridges erected should 
be made, first, for the next greater price and then for the next lower 
one, after which the necessary interpolation would be a simple matter. 

The proportion of the total cost of the erected metal which pertains 
to the erection has been arbitrarily assumed for convenience at 20 per 
cent. This gives a range of from 0.5 cent to 1.2 cents per lb. as the 
cost of erection, which is a fair assumption, for, even in elevated-rail- 
road work, the cost of erection and painting is seldom as low as h cent 
per lb., and for railroad bridges, even in remote localities, it does not 



148 NICKEL STEEL FOR BRIDGES 

often exceed 1.2 cents per lb. For cantilever bridges, the cost of erec- 
tion might be more, and thus a comparison of cost slightly too favor- 
able to nickel steel might be made, were it not for the fact that in 
such large structures the use of this metal will have a tendency to 
lower comparatively the pound cost of erection, because the decrease 
in weight of individual members facilitates progress and reduces the 
cost of the traveler, derricks, and other heavy apparatus. On the 
whole, the assumption made for the proportionate cost of erection is, 
perhaps, the fairest that could be adopted. 

In comparing the costs of erection of nickel-steel bridges and 
carbon-steel bridges due cognizance was taken of the fact that, for two 
similar bridges of equal carrying capacity, while the total cost of 
erection of the nickel-steel structure is less than that for the carbon- 
steel structure, the cost per pound in the former is greater than in the 
latter, because certain items of expense are constant while others vary 
with the weight of metal handled. The writer has assumed that one- 
half the total expense is constant and that the other half will vary 
directly with the weight of metal. This is as accurate a division as 
can be assumed. Upon this basis was established the following mathe- 
matical statement: 

Let W = weight of metal per linear foot of span in the 
carbon-steel bridge, 
W = ditto for the nickel-steel bridge, 
C = cost per pound for erecting the carbon-steel 

bridge, 
C" = ditto for the nickel-steel bridge, 
F = cost per linear foot for erecting the nickel-steel 
bridge, 
then CW = cost per linear foot for erecting the carbon-steel 
bridge, 
C W 

9 



-=^"(' + D 



In plotting the curves of cost of nickel-steel bridges and mixed- 
steel bridges erected, given on the diagrams, the cost per pound of the 
erection of the alloy was computed by this last equation. 



NICKEL STEEL FOR BRIDGES 



149 



The types of bridges covered by the diagrams are as follows: 
Single-track, deck, plate-girder spans. 
Single-track, half-through, plate-girder spans, 
Single-track, through, riveted, Pratt-truss spans. 
Single-track, through, pin-connected, Pratt-truss spans, 
Single-track, through, pin-connected. Petit-truss spans. 
Double-track, through, riveted, Pratt-truss spans. 
Double-track, through, pin-connected, Pratt-truss spans. 
Double-track, through, pin-connected. Petit-truss spans. 
Double-track, through, piu-connected, "Type A^' cantilever 

bridges. 
Double-track, through, pin-connected, "Type i?," cantilever 

bridges. 
Double-track, through, pin-connected, "Type C," cantilever 

bridges. 
Double-track, through, pin-connected, "Type D," cantilever 

bridges. 



These twelve types (barring double-track, plate-girder spans) cover 
practically all the bridges that are built nowadays in the United States. 
In the case of any type not included in the preceding list, the dia- 
grams may be used by adopting that one for the structure most like 
it and for the existing conditions of the metal market. The differences 
between Types A, B, G, and D of cantilever bridges, illustrated on 
Fig. 22, were explained previously. In the diagrams of weights of 
these cantilever bridges it must not be forgotten that the weights of 
metal per linear foot of bridge given are the averages from end to end 
of structure, and are not the weights per foot for any particular span 
or spans. 

In computing the weights of metal for various spans which are 
plotted on Figs. 10 to 21, inclusive, it was found, as might readily 
have been anticipated, that the economic truss and girder depths are 
somewhat less for nickel-steel than for carbon-steel bridges. The rea- 
sons for this are: first, that, in comparison with carbon steel, nickel 
steel can be strained higher in the compression members of chords 
than in those of webs; and, second, that in the webs of nickel-steel 



150 NICKEL STEEL FOR BRIDGES 

bridges there are necessarily more minimum sections used than in 
those of carbon-steel bridges. 

It is evident that, on account of both the smaller economic depths 
and the higher intensities of working stresses, the deflections of nickel- 
steel spans will be greater than those of corresponding carbon-steel 
spans. However, this increase of deflection is not a matter of any 
great importance. 

If, eventually, nickel steel should supplant carbon steel in bridge- 
work, the latter metal will continue to be used for a long time in parts 
that do not take direct stress (such as stay-plates, lacing bars, and web 
stiffeners), in the lateral systems of all bridges, except those of ex- 
tremely long spans; and in truss members having much larger sec- 
tions than the stresses call for, such as web members near mid-span 
and the secondary vertical posts of Petit trusses. 

In some cases, especially when bridge metal is cheap, a still fur- 
ther saving might be effected by making the entire floor system of 
carbon steel; but as the amount of money thus gained in the floor 
would be small, and as it would have to be reduced somewhat by the 
increased cost of trusses due to the slightly greater dead load, this 
kind of economy is problematical. In long-span bridges the necessity 
of keeping the dead load as low as possible would preclude the adoption 
of carbon steel for the floor system, even if the use of such steel there 
were per se decidedly the cheaper. 

In all the diagrams it has been assumed that carbon steel would 
be used exclusively for lateral systems; but it is a fact that in long 
spans it would be economical to adopt nickel steel for some of the 
heavier lateral members, consequently, the comparative costs of long- 
span bridges, of carbon-steel and of mixed-steel, in case of actual de- 
signs with careful detailing, might show even greater differences than 
those given by the curves. 

In order to demonstrate how the diagrams are to be used, it will 
be well to assume a few cases and apply the curves to their solution. 

Case 1. — A long, single-track bridge consists of a succession of half- 
through, plate-girder spans of 100 ft. each, carbon steel erected cost- 
ing 4.5 cents per lb., and nickel steel delivered at site being worth 1.6 
cents per lb. more than carbon steel. Find the comparative costs of 
carbon-steel and mixed-steel bridges. 



NICKIiL STEEL FOR BRIDGES 151 

Turning to Fig. 31, there are found $101.50 as "the cost per linear 
foot of the metal in the carbon-steel bridge, and $92 as the correspond- 
ing cost for the mixed-steel bridge. 

Case 2. — A double-track bridge consists of four riveted, through 
spans of 200 ft. each, carbon steel erected costing 4 cents per lb., and 
nickel steel delivered at site being worth 1.4 cents per lb. more than 
carbon steel. Find the comparative costs of carbon-steel and mixed- 
steel bridges. 

Turning to Fig. 51, there are found $188 as the cost per linear 
foot of the metal in the carbon-steel bridge, and $174 as the corre- 
sponding cost for the mixed-steel bridge. 

Case S. — A double-track. Type A, cantilever bridge has a main 
span of 1 050 ft. If built of carbon steel, it would cost 5.5 cents per 
lb. erected. Nickel steel delivered at site is worth 1.5 cents per lb. 
more than carbon steel. Find the comparative costs of carbon-steel 
and mixed-steel bridges. 

Turning to Fig. 69, there arc found the following: 

Carbon-steel bridge $658 

Mixed-steel bridge for excess of 1.6c 555 

" " " for excess of 1.4c 543 

By interpolation for excess of 1.5c 549 

An attempt will be made to anticipate what will be the probable 
excess pound price for shopwork on nickel steel as compared with car- 
bon steel, using as a basis the present average cost of reamed shop- 
work, which is approximately to cent per lb. in the principal Ameri- 
can bridge shops. This figure includes all expenses and fixed charges 
of every kind, such as heat, light, power, and office expenses. 

To obtain a result that is closely accurate, it will be necessary to 
itemize the various shop costs for carbon-steel work, add to each item 
its pro rata share of general expense, determine for each item the 
approximate ratio of increase for nickel-steel work, and calculate the 
increased items and increased total cost per pound. 

The division of cost given in Table 5 is probably as good an average 
as could be assumed; but it mvist be remembered that any division 
whatsoever would vary with the style and individuality of the shops, 



152 



NICKEL STEEL FOR BRIDGES 



with the character of the construction, and even with the personnel 
of the shop management. Table 5 gives all the information required 
for obtaining the cost of reamed shopwork on nickel steel for bridges. 

Each of the ratios of increased cost given in Table 5 was carefully 
considered in consultation with John Lyle Harrington, M. Am. Soc. 
C. E., and the items of division of shop cost were furnished by him. 

Table 5 shows that the excess cost per pound for the manufacture 
of nickel-steel bridges, as compared with carbon-steel bridges, is 0.15 
cent. Curiously enough, this is exactly the figure named to the writer 
four years ago as an off-hand guess by C. C. Schneider, Past-Presi- 
dent, Am. Soc. C. E. 

TABLE 5. — Comparison of Cost of Shopwork per Pound fob 
Carbon Steel and Nickel Steel. 



Items. 



Drawing-room work 

Template-shop work 

Laying- ovit work 

Shearing and straightening 

Punching 

Assembling and bolting 

Reaming and di-illing 

Chipping and milling 

Riveting 

Painting 

Miscellaneous 

Total and average 



Cost of shop- 
work per pound 
for carbon 
steel. 



0.08 cent. 

0.04 

0.04 

0.04 

0.08 

0.12 

0.15 

o.oa 

O.IG 
0.03 
0.04 



0.80 cent. 



Ratio of 

increased cost 

for nickel 

steelwork. 



1.25 
1.25 
1.10 
1.10 
1.35 
1.10 
1.10 
1.50 
1..30 
1.25 
1.00 



1.19 



Cost of shop- 
work per pound 
for nickel 
steel. 



0.100 cent. 

0.050 " 

0.044 " 

0.044 " 

0.100 " 

0.132 '• 

0.165 " 

0.030 " 

0.208 '• 

0.037 " 

0.040 " 



0.950 cent. 



Nickel-steel ingots, when nickel is worth 30 cents per lb., cost 
about 1 cent per lb. more than those of carbon-steel; and the differ- 
ence in cost of rolling should certainly not exceed tit cent per lb. 
Allowing 20% profit on these excess costs would make the total excess 
cost of nickel-steel bridgework delivered at site 1.5 cents per lb. The 
almost exact agreement of this difference with that for the eye-bars 
in the Manhattan Bridge appears to be pretty conclusive. 

Adopting, then, 1.5 cents as the probable difference in pound prices 
of nickel steelwork and carbon steelwork delivered at site, it will be 



NICKEL STEEL lOK BRIDGES 



153 



interesting to compare the costs of bridges of carbon-steel and of 
mixed-steel of all the twelve kinds covered by the diagrams. 

In making this comparison, it will be assumed that the average 
pound prices for carbon-steel bridges erected throughout the United 
States are as follows: 

Plate-girder spans 4.0 cents. 

Riveted-truss spans 4.5 " 

Pin-connected, Pratt-truss spans 4.5 " 

Pin-connected, Petit-truss spans 5.0 " 

Cantilever bridges 5.5 " 

The reason for the greater assumed pound costs of long-span 
bridges is mainly expensive erection, because such spans are generally 
used where the erection conditions are costly. 

TABLE 6. — Percentages of Excess of Cost of Carbon-Steel 
Bridges over Mixed-Steel Bridges. 



Type of Structure. 



Single-track, deck, plate-girder spans 

Single-track, lialf-througli, plate-gii'der spans 

Singlt'-track, through, riveted, Pratt-truss 
spans 

Single-track, through, pincounected, Pratt- 
truss spans 

Single-track, through, pin-connectsd, Petit- 
truss spans 

Double-track, through, riveted, Pratt-truss 
spans 

Double-track, through, pin-connected, Pratt- 
truss spans 

Double-track, through, pin-connected, Petit- 
truss spans 

Cantilever bridges of " Type .4 " 

Cantilever bridges of " Type i? " 

Cantilever bridges of " Typ» (' ' 

Cantilever bridges of '■ Type Z> "' 

General average for all bridges 



Least. 



— 5 

+ 3 

+ 2 
+ 1 
-f 10 
+ i 
+ 2 

+ 13 

f 7 
9 



I 



-1-5 



Greatest. 



-f 11 

+ 12 



+ 11 

+ ir 
+ « 
+ 11 

4-20 
-f-30 
-i-29 
-f25 
+ 26 



+ 17 



Approximate 
average. 



+ 5 

+ '< 

+ 5 

+ 

+ 14 

+ 6 

+ (5 

16 
18 
19 
+ 18 
+ 20 



t 



+ 12 



From Fig. 25 it is found that for deck, plate-girder spans, carbon- 
steel bridges are cheaper than mixed-steel bridges only for spans of 
less than 33 ft., and that in all greater spans they are more expensive. 



154 NICKEL STEEL FOR BRIDGES 

A study of Figs. 30, 3G, 41, 47, 52, 57, 63, and 69 shows that, for the 
conditions assumed, carbon-steel bridges are invariably more expensive 
than those of mixed nickel and carbon steels. The percentages of the 
greater cost are given in Table 6. 

Summarizing, it is evident that it would be economical at the pres- 
ent time to use nickel steel for all kinds of railroad bridges, and the 
longer the spans the greater the economy. It might be shown also 
that nickel steel would be economical for certain highway bridges, but 
its adoption would certainly be inadvisable for ordinary county bridges, 
because the use of the new metal might cause such structures to "vanish 
into thin air." 

The general use of nickel steel for bridges not only would result 
in decidedly cheaper structures, but also would permit of the building 
of longer spans than are at present attainable. 

For instance, it is generally conceded by bridge engineers that the 
present greatest practicable main-span length for cantilevers built of 
carbon steel is in the neighborhood of 2 000 ft. On Fig. 71 are plotted 
the probable weights of metal per linear foot of bridge in carbon steel 
and in nickel steel for main spans far longer than any yet designed or 
computed. The method adopted for plotting was to record the already 
diagrammed weights for spans of 1 200 ft., 1 350 ft, 1 500 ft., 1 650 ft., 
and 1 800 ft., pass through the five points of each record a circular 
curve, and carry that curve to the limits of the paper. The weights 
of metal thus established for spans of unprecedented length are prob- 
ably fairly accurate. In any case they are sufficiently so for the pres- 
ent purpose, which is simply to determine approximately the main span 
lengths of cantilever bridges of "Type A" that have the same average 
weight of metal per linear foot for the two kinds of steel. The dia- 
gram gives the following corresponding main-span lengths : 

Carbon-steel bridges 1 300 ft . 1 400 ft. 1 (500 ft. 1 800 ft. 2 000 ft . 

Nickel-steel bridges 1050" 1830" 2 040" 2 300" 2 60)" 

From this it will be seen that if 1 800 ft. be assumed as the present 
practicable limit of span length for carbon-steel bridges, the corre- 
sponding limit for nickol-stcel bridges will be about 2 300 ft.; or, if 
it be assumed at 2 000 ft., the corresponding limit for nickel-steel 
construction will be 2 600 ft. It is safe, therefore, to conclude that 
the adoption of nickel steel for bridges would lengthen the practicable 



NICKEL STEEL FOR BRIDGES 155 

span length for cantilevers fully 500 ft. The writer foretold this re- 
sult just before the experiments on nickel steel for bridges were in- 
augurated. 

If the question of greatest span leng-th be one of economics instead 
of practicability, the curves on Fig. 72 should be used. These were 
prepared from the data on Fig. G9, assuming that the excess value of 
nickel steel delivered at site is 1.5 cents per lb. 

The diagram gives the following corresponding lengths of main 
spans for equal costs per linear foot of bridge: 

Carbon-steel bridges I'SOO ft. 1 400 ft. 1 GOO ft. 1 800 ft. 2 000 ft. 

Nickel-steel bridges r40(i " 1 GOO '• 1 cSOO " 2 010 " 2 2G5 " 

The principal application of the results of this last investigation 
is to the comparative economy of cantilever and suspension bridges; 
for if it be known that for carbon-steel construction the length of 
cantilever main span corresponding to equal cost per foot be 2 000 ft., 
it can be seen from the table that for nickel-steel construction it is 
2 265 ft. 

Summarizing the results of this entire investigation of nickel steel 
for bridges, it is evident that nickel steel is in every way fitted for 
bridge construction, in that it is strong, tough, workable, and reliable; 
moreover, its adoption would effect a decided economy. This economy 
would increase in the future as the cost of nickel decreases and as the 
shops become more accustomed to the fabrication of the new alloy. 
That nickel will soon be less expensive is a foregone conclusion, in view 
of the immense deposits of nickel ore that have been located and sur- 
veyed in Canada. It is said upon good authority that there has been 
found in one deposit in that country ore containing fully 200 000 tons 
of the metal. 

Wliile the writer has never known nickel to have been sold for less 
than 30 cents per lb., nevertheless, he is of the opinion that, should this 
material be called for in large tonnages for bridge building, it might 
be purchased as low as 25 cents. It makes a great difference in the 
price to the producer whether a metal is sold by the pound or by the 
ton ; and tons of nickel would be required where pounds are bought to- 
day, were nickel steel used extensively for bridgework. 

At 25 cents per lb. for nickel, the price of rolled nickel steel would 
be about 0.2 cent per lb. lower than it would be with nickel at 30 cents 



156 NICKEL STEEL EOR BRIDGES 

per lb., and it is likely that the cost of manufacture would be reduced 
in the same proportion, thus making the price of nickel-steel bridge 
metal delivered at site 1.2 cents per lb. above that of carbon steelwork. 
A study of the accompanying diagrams will show the great economy of 
using nickel steel for bridges under such conditions. 

In concluding this paper the writer desires to ask for a thorough 
discussion, and to express the hope that the effect of the paper and 
the discussion will be to hasten materially the adoption of nickel steel 
for bridges. 



NICKEL STEEL FOR BRIDGES 



157 




=^i=-4-# 







05 


oc 




bo 


m 
o 


G ■ 




bo 


H 


( ) 




Q-> 


O 




[J 


y^\ 


^ 






V y 


'^ 


o 


1 


<j 


t; 






')0 

po 




J 



^.1 




158 



NICKEL STEEL FOR'^ BRIDGES 



J 




12 3 4 5 

Siiuare of Deflection, in Inches 
Fig. 3. 



NICKEL ^TEEL FOR BRIDGES 



159 



TESTS OF RELATIVE CORROSION 

OF CARBON AND NICKEL STEELS. 

SULPHURIC ACID. 



Uy The Osborn Kngincering Co. 
(Oue Per Cent Solution) 



By J.A.L.Waddell. 
(Two Per Cent Solution) 




20 40 60 80 , 100 

Loss of WeigM, in Percentage. 
Fig. 4. 



IfiO 



NICKEL STEEL FOR -BRIDGES 



TESTS OF RELATIVE CORROSION 
OF CARBON AND NICKEL STEELS. 



i!y Tlic Osborn Engineering Co. 



By J.A.L.Waddell 




550 



500 



450 



400 



350 « 

e3 
Q 

C 

300 II- 

■s: 
o 
H 

o 
250 PI 



200 p 



150 



100 



50 



1 



3 0.5 1.0 

Loss of Weight, in Perccntago 

Fio. 5. 



1.5 2.0 



NICKEL STEEL FOR BRIDGES 



161 




12 3 4 

Loss of Weight, in Percentage 

FiQ. 



162 



NICKEL STEEL FOR BRIDGES 



.TESTS OF RELATIVE CORROSION 
OF CARBON AND NICKEL STEELS. 



By The Osborn Engineering Co. 



ByJ.A.L.Waddell. 




15 20 1 

Loss of Weight, in Percentage, 

Fig. 7. 



NICKEL STEEL FOR BKIDGES 



163 




Fig. 8. 



164 



NICKEL STEEL FOR BRIDGES 



Load, in IncliPound Moments. 




NICKEL STEEL FOR BRIDGES 



1G5 



IGOO 



liipi#[H:|tl i-H f HJ:| 1 1 !i!t:Hf |#tiHv[+|iiimHj#H4tttfttftttH#(H l\im 



loOO 



1 :m 

1 -,'00 

i 1 100 

in 

o 

■g 1 000 • 



WtiuriTcj uf- 
SINGLE-TRACK, DECK, 
PLATE-GIRDER SPANS. 
LIVE LOAD- CLASS R. 



loot 



Xoti': To ol)Laiii the wt'jght of carbon steel 
in spans where both steels are used, 
take the difference in the readings |; 
given by the two lower lines. 



A L 



— t: 



IGOO 



1 200 




200 



20 30 40 jO 00 iU >0 00 100 110 

Span, in Feet. 

Fig. 10. 



100 



166 



NICKEL STEEL FOR BRIDGES 



WEIGHTS OF 

SINGLE-TRACK, HALF-THROUGH, 

PLATE-GIRDER SPANS. 

LIVE LOAD, CLASS R. 




200 



Note: To obtain the woiglit ol" carbon steel 
in spans where both steels lire used 
take the (lill'erriue in the readings 
given liy the two lower lines. 




M-l-ltil4-itimu 



2 000 



2 400 



2O00 



1 800 



1 000 



1 100 



1 200 



800 



000 



100 



30 40 50 60 rO 8 

Span, in Feet 
Fig. 11. 



90 100 110 120 



200 



NICKEL STEEL FOR BRIDGES 



167 



WEIGHTS OF 

SINGLE-TRACK, THROUGH, RIVETED, 

PRATT-TRUSS SPANS- 

LIVE LOAD, CLASS R. 



2 600 
2 500 
2 400 
2 300 
2300 
2 100 
2 000 
1 900 
1 800 
1 TOO 
1 000 
1 500 
1 -400 

1 .-^oo 

1 200 
1 100 
1 000 
900 
800 
TOO 
GOO 





100 



Ncte: To obtain the weight of carbon .slii 1 
in spans: where both steels are used, 
talve the dilierence in the readings 
given by tlie two lower lines. 



2 600 
2 500 
2 400 
2 300 
2200 

2 100 
2 000 
1 900 
1 800 
1 TOO 
1 COO 
1 500 
1 400 

1 300 

1 200 

1 100 

1 000 

900 

800 

700 

GOO 



110 



120 



T-BTffnffffffFfffffl ilH 



130 



^ffrk 



140 150 160 

Span, in Feet 
Fig. 13. 



ITO 



ISO 



.190 



200 



168 



NICKEL STEEL FOR BRIDGES 



3 800 



3 400 



3 200 



3 000 



« 2 800 

m 



2 600 



-r 2 200 



.= 2 000 



M 



> 1800 



3 600 



1400 



1200 



1000 



WEIGHTS OF irf 

SINGLE-TRACK, THROUGH, PIN-CONNECTED, i^ 
PRATT-TRUSS SPANS '*" 

LIVE LOAD CLASS R 



800 



+t+l-H-H-i- 



-l4ffl+i-[ 



Sffi 




4-t[^IUI-Ff+l4-tJ44l]-l 



Note: To obtain the wciglil of carbon stool 
ill spans ulioie both stools are tised, 
take' the diU'oronce iu the readings I 
givoii by the two lower Hues 




3 000 



3 100 



3 200 



- 3 000 



2 800 



2 600 



2 400 



2 200 



>000 



1200 



1000 



irO 190 210 230 



250 2r0 290 

Span, in Feet 

Fig. 13. 



310 330 350 370 



800 



NICKEL STEEL FOR BRIDGES 



169 



0500 




0500 



4 900 



4500 



4 100 



3 700 



5 500 



~'100 



i;oo 



1.3U0 



900 



350 375 400 425 450 475 500 525 550 575 000 

Span, in Feet 
Fig. 14. 



500 



170 



NICKEL STEEL FOR BRIDGES 



5000 



laSOOO 




2 000 



100 110 120 130 140 150 100 170 180 190 200 

Span, in Feet 
Fig. 15. 



NICKEL STEEL FOR BRIDGES 



171 



6600 



OCOO 



000 



700 




3 400 



2 1001^^ 



2 4D0 



180 200 220 240 260 280 300 320 340 360 380 

Span, in Feet 
Fig. 16, 



2 100 



172 



NICKEL STKKL FOR BRIDGES 



10 500 



10 000 



9 500 



10 500 



10 000 



J 9 500 




425 m 475 500 

Span, in Feet. 

Fig. 17. 



NICKEL STEEL FOR BRIDGES 



173 



30 000 



28 000 



20 000 



24 000 



22 000 



20 000 



IS 000 



16 000 



14 000 



WEIGHTS OF 
DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 




J+l-H-IIIIIIII Trl f-MliP-Kl-i-HtHH-f 



CANTILEVER BRIDGES OF 
TYPE .1. 

Live Load-Class R for Stringers, 
Class S for Floor Beams and 
Pi iniary Trus', Menibei s, and 
Class I for Main Truss ^lembers 



S 12 000 



f: 10 000 
*S 

8000 



GOOC 



4 000; 



2 000: 




fern 



iiTi i itii l ii in i nH fmtH-tl 



Xotc: To obtain the weight of carbon steel 
in spans wlieie both steels are used, 
take tlic ciillerence in the readings 
uivtMi Ii\" tlic two lower linos. 



MO 000 



•2S 000 



■,^i 000 



24 000 



')2 000 



20 000 



I b 000 



11000 



1 i 000 



GOOO 



4 000 



;)000 



300 450 600 750 900 1050 1200 1350 

Length of Main Opening, in Feet. 

Fm. 18. 



174 



NICKEL STEEL FOR BRIDGES 



30 000 



28000 



26 000 



-Si- 



24 000 



WEIGHTS OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 

CANTILEVER BRIDGES OF 

TYPE B 

Live Load— Class 11 for Stringers, 

Class S for Floor Beams and 

Primary Truss Members, and 

Class U tor Main Truss Members. 




30 000 



38 000 



20 000 



24 000 



22 000 



20 000 



18 000 



14 000 



10 000 



8 000 



000 



4 000 



1000 



750 900 1 050 1 200 1 350 

Length of Main Opejung, in Feet 

Fig. 19. 



1800 



NICKEL STEEL FOR BRIDGES 



175 



WEIGHTS OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 

CANTILEVER BRIDGES OF 

TYPE C 

Live Loa(l-Cl;iss R for Stringers, 

Class S for Floor liearas and Primary Truss Meniljers, 

anrl Class U for Main Truss iVIembers 



24 000 



22000 




24 000 



20 000 



IS 000 



16 000 



uooo 



12 000 



10 000 



1000 



cooo 



2 000 



750 900 1 050 1 200 1 350 1 500 1 650 1 800 

Length of Main Opening, in Feet 
Fig. 20. 



176 



NICKEL STEEL FOR BRIDGES 



30 000 



28 000 fe 




30 000 



.'5 000 



30 000 



34 OUU 



33 000 



000 



18 000 



lie 000 



^- 1 1 OOO 



13 000 



6 000 



300 450 COO 750 900 1 uoo ) -M) 1 350 1 500 1 650 1 800 

Length of ^lain Opening, in Feet 



NICKEL STEEL FOR BRIDGES 



177 



DOUBLE-TRACK RAILROAD CANTILEVER BRIDGES. 
TYPICAL LAY-OUTS. 

Trusses are spaced to agree with requirements of "De PontiljusV 
Loading is Class R for stringers, Class S for floor beams and 
primary truss meinl)ers, and Class U for main truss members. 
Suspended spans are >» of main opening and cantilever and 
anchor arms are each *io of main opening. 
I equals sum of one suspended span and two cantilc\er arms. 




Fig. 22. 



178 



NICKEL STEEL FOR BRIDGES 




l;i.i){) 



;iU to .".0 GU :u SU 'M UNi lU) 

Span, in Feet 
Figs. 23 and 24. 



NICKEL STEEL FOU BRIDGES 



179 



OS 



Gl 



00 



5C 



48 




COMPARATIVE COST OF 

SINGLE-TRACK, DECK, PLATE-GIRDER 

SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS 
CARBON, STEEL ERECTED COSTING 

4.0 CENTS PER POUND 



Kickci Stui 
Ni.jkul St.;L-l, 



CARBON 



I.S .■iiith 



MclL4 StL-L-1, 1.1'. rrlit., K' 

I I 

Nickel Strcl, 1.1 .-at, Kx. 
.N'lL'kul Steel, 1.2 eelit, K.x. II. 



Xiekel StJIl, l.iJo-iit' Ivv D.i; 

^i^kc\ Steel 0.^ cel.l Iv\. O'-l. 
Nlikel Steel, U.f. eent 1",.\. U.-l. 




40 - 



a 3G 



» 32 

o 

a 28 



12 




68 



CO 



40 



32 



28 



24 



20 30 40 50 60 70 

Span, in Feet 
Fig. 25. 



80 90 100 110 



180 



NICKEL STEEL FOR BRIDGES 




50 CO VO 

Span, in Feet. 

FuiS. 2(1 AND 37. 



NICKEL STEEL FOR BRIDGES 



181 



CuMPAHATIVE COST OF 
SINGLE-TRACK, HALF-THROUGH, PLATE-GIRDER 
bPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

3.0 CENTS PErt POUND 




21 



20 






rrnj^-ri-UiUHM+Hm- 



30 



40 



CO ro ( 

Span, in Feet 
Fig. 28. 



90 



ICO 110 120 



183 



NICKEL STEEL FOR BRIDGES 




60 70 80 

Span, in Feet 



90 100 110 120 



Fig. 20. 



NlCKIili STEliL FOll BRIDGES 



183 




IP- 



Jll-IUII.IIli-l|-t-lHllll|i;i! 11:1 



SINGL 



COMPARATIVE COST OF 
E-TRACK, HALF-THROUGH, PLATE-GIRDER 

SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

4.0 CENTS PER POUND 



IMmWilSil 



il05 




60 70 80 

Span, in Feet 

Fig. 30. 



184 



NICKEL STEEL FOR BRIDGES 




30 10 OU 



Span, in Feet 
Fig. 31. 



lUU 110 VM 



NICKEL STEEL FOR BRIDGES 



185 




Span, in Feet 
Fig. 32. 



186 



MICKEL STEEL FOR BRIDGES 




140 150 ICO 

Span, in Feet. 
Figs. .33 and .34. 



NICKEL STEEL TOll BUIUOES 




COMPARATIVE COST OF 

SINGLE-TRACK, THROUGH, RIVETED, PRATT-TRUSS 

SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

4.5 CENTS PER POUND 



140 150 160 

Span, in Feet. 

Figs. 35 and 36. 



188 



NICKEL STEEL FOR BRIDGES 




100 110 120 ISO 140 150 160 170 180 190 

Span, in Feet 

Fig. 37. 



NICKEL STEEL FOR BRIDGES 



189 



130 



1^0 



i 



100 



<» 80 



80 



COMPARATIVE COST OF 

SINGLE-TRACK, THROUGH, PIN-CONNECTED, 

PRATT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

3.0 CEN- 



impHHSfffiHreTj 





COMPARATIVE COST OF 
SINGLE-TRACK, THROUGH, PIN-CONNECTED, 
PRATT -TRUSS SPANS OF CARBON STEEL AND 

MIXED 

:" NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

3.5 CENTS PER POUND 



120 



110 



100 



SO 



GO 



50 



■Zoi) 2^0 -ZM 

Spun, in Feet. 
Figs. 38 and 39. 



.};o 



190 



NICKEL STEEL FOR BRIDGES 



160 



COMPARATIVE COST OF 
SINGLE-TRACK, THROUGH, PIN-CONNECTED, 
PRATT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

4.0 CENTS PER POUND 




80 



>< 



COMPARATIVE COST OF 

SINGLE-TRACK, THROUGH PIN CONNECTED, 

PRATT-'l RUSS SPANS OF CARBON STEEL ANC 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

4.5 CENTS PER POUND 

: I I I :'m 



190 210 



330 



250 270 290 

Span, in Feet. 

Figs. 40 and 41. 



310 



330 




NICKEL STEEL FOR BRIDGES 



191 




90 
ITO 



COMPARATIVE COST OF 

SINGLE-TRACK, THROUGH, PIN-CONNECTED, 

PRATT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

5.5 CENTS PER POUND 



no 



T n lfi M 1 ii' ill i i 



190 



;J10 




250 270 290 

Span, in Feet. 

Figs. 43 and 4.3. 




192 



NICKEL STEEL FOR BRIDGES 



350 LM j 1 1 1 u ij 1 1 1 1 1 1 1 1 


::::x::: ::t: :::-::::! 


; 1 1 1 1 1 1 t 1 1 1 1 ) 1 I 1 


TTl 


ITTT 






TP 


fTTT 




250 




::.-.+ X J T 


± 


--- 


"-- 








--j-_ 








X i 


± __t: — 


— 









-y 


.^_ 




















































— ^" ' ^"=." T r ^fl="""# 


+1 — h+i 1 — 


-1^- 


-X 


^^^ 


™J- 


—1- 


f-—^- 


-p-L 




jj40::j#::::JUJJJl 


COMPARATIVE COST OF 

LE-TRACK, THROUGH, PIN-CONNEC 

T-TRUSS SPANS OF CARBON STEEL 

MIXED 

NICKEL AND CARBON STEELS. 

CARBON STEEL ERECTED COSTINC 


TPn ■- — 


t^ 


^\ 


-x- 


:- :•.-.- 


Si40 


:::::--:::' PET 


TED, 

AND 






330 


230":"":-p" 








22u 


3.5 CENTS PER POUND 

ft 








230 










y 


•'10 




flfffH 






/- y 




■#rmU 


fiffi: 


/ 




/^ 


X 


Z 


300 






/ . 


X- - ■ - 


y 

y- ■ • - y 


190 X~ ^— 

ISO 


x/ X. 




X 


T^y 

y 


190 


iro 


xy{^X> 


Xs>x^^X "i^ y^ 

, ■■■" <.~ " ^\ y 


^' y 

y 

y ^ 

Xyyy^ 


iro 


X ' 
IGO 

15U ^ 

140 ''XX—tXX^ 


^X:, y y 


^y 
' ■■X 


V ., ^ ;- 


yxx- 


100 
120 
Ul-t 


1-^0 ^XXr~X^ 

110 ^--U=-tr- J 

X-, --XX 


X 1 :.L - XX-- 

----_-----------xX-'-xxxx''\'''' 


||^t^#t|-:, 


. 




— 


130 

13U 
110 


lUO it :....:-r.. 


:;:: "---:" j ' " " :"" ;/ -. ., ! 


--■■-llll, ililjilil 


iliiiWii 


1 . . , 




100 



350 375 400 425 450 475 500 5^ 550 575 UOJ 

Span, in Feet 
Fig. 44. 



NICKLir. STIiEL I'OU BRIDGES 



193 




450 475 500 

Span, ill Feet 
Fig. 45. 



194 



NICKEL STEEL FOE BRIDGES 



300 



COMPARATIVE COST OF 
SINGLE-TRACK, THROUGH, PIN-CONNECTED, 
PETIT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

4.5 CENTS PER POUND 






300 



m 



^u 




450 475 500 

Span, in Feet. 
FlOS. 46 AND 47, 



NICKEL STEEL FOR BRIDGES 



195 

— 1 100 



COMPARATIVE COST OF 
SINGLE-TRACK, THROUGH, PIN-CONNECTED, 
PETIT-TRUSS SPANS OF CARBON STEEL AND i 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

5.5 CENTS PER POUND 



3S0 




450 47o 500 

Span, in Feet 



Fig. 48. 



196 



NICKEL STEEL TOR BRIDGES 




100 110 



190 



130 140 150 160 170 

Span, In Feet 

Figs. 49 and 50. 



190 800 



NICKEL STEEL FOi; I'.KIDCES 




100 110 v^o 



lao I-IO 150 160 

Span, in Feet 

Figs. 51 and 53. 



198 



NICKEL STEEL FOR BRIDGES 



250 



iMO 



xffl 



fflTHTr-r 



COMPARATIVE COST OF 
DOUBLE-TRACK, THROUGH, RIVETED, PRATT-TRUSS 
SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

5 CENTS PER POUND 



»'W 




140 150 100 

Span, in Feet 

Fig. 53, 



NICKEL STEEL FOR BRIDGES 



199 



COMPARATIVE COST Of 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, PRATT 

TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

3.0 CENTS PER POUND 




-••.'0 




100 



f "--T 



1-^0 



^-r H- 



d_ 



:so 



180 



200 



220 



210 



260 280 300 

Span, in Feet 
Fio. 54. 



M) 



360 



380 



200 



NICKEL STEEL FOR BRIDGES 




a 190 
2 



COMPARATIVE COST OF 
DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 
PRATT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

3.5 CENTS PER POUND 




200 280 300 

Span, in Feet. 

- Fig. 55. 



NICKKL STEEL FOIt 15i;il)(lES 



201 



^o:..; ;,...J iVE COST OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 
PRATT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

4.0 CENTS PER POUND. 



s) 2:50 




260 280 300 

Span, in Feet 

Fia. 56. 



202 



NICKEL STEEL FOR BRIDGES 




■^(iO ^'80 300 

Span, in Feet 

Fig. 57. 



NICKEL STEEL FOR BRIDGES 



203 

H-H-ffl-l+f H-h '" 1*^ 




180 200 



Span, in Feet 
Figs. 58 and 59. 



204 



NICKEL STEEL FOR BRIDGES 




220 



200 



1 



COMPARATIVE COST OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 

PETIT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

10 CENTS PER POUND 



240 



SCO 



ilMMIMi 



200 



350 375 



450 475 500 

Span, in Feet 
Figs. 60 and (il. 



550 



000 



NICKEL STEEL FOR BRIDGES 



205 




450 475 500 

Span, in Feet 

Fig. 62. 



206 



NICKEL STEEL FOR BRIDGES 



5i0 



, ST." 



COMPARATIVE COST OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 

PETIT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 

CARBON STEEL ERECTED COSTING 

5.0 CENTS PER POUND. 



ri540 




3401- 



Hiili-riiiMffiftH 



260 



350 



375 400 425 



150 475 500 

Span, in Feet 

Fig. 63, 



5-35 550 575 000 



240 



NICKEL STEEL FOR BRIDGES 



207 



r)80 




COMPARATIVE COST OF 

DOUBLE-TRACK, THROUGH, PIN-CONNECTED, 

PETIT-TRUSS SPANS OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

5.5 CENTS PER POUND 




450 475 500 

Span, in Feet 



Fig. 64. 



NICKEL STEEL FOR BRIDGES 



1100 



1000 




900 1 050 1 200 1 350 

Lcugtli of Main Oiiening, iu Feet 
Figs. 05 and 66. 



100 
1800 



NICKEL STEEL FOR BRIDGES 




COMPARATIVE COST Of 
DOUBLE-TRACK, THROUGH PIN-CONNECTED, TYPE.4 
CANTILEVER BRIDGES OF CARBON STEEL AND 

MIXED 

NICKEL AND CARBON STEELS, 
CARBON STEEL ERECTED COSTING 

5.0 CENTS PER POUND 



300 



600 



750 



uoo 



1030 



1200 1350 1500 1050 1800 



Length of Main Opening, in Feet, 
Figs. 67 and 68. 



210 



NICKEL STEEL FOE BRIDGES 



1600 




1600 



1500 



300 450 oOO 750 'M> 



1 200 1 350 1 500 1 050 1 SOO 



Length oi .\i.iin 1 'pc-iiing, in Feet 
Fig. go. 



NICKEL STEEL FOR BRIDGES 



211 



1000 




300 4j0 



750 900 1 050 1 200 1 350 1 V 

Length of Main Opening, in Fent 
Fig. 70. 



213 



NICKEL STEEL FOR BRIDGES 



2 700 




2 700 



2 600 



500 



400 



2 300 



200 



2 100 



2 000 ^ 



1 1100 E 



1 800 



1 ;oo 



1 uoo 



1 ."A)0 



1 100 



- 1 ;ioo 



1 200 

3ii 000 'S> uoo li-j uiio xi'.i oiiii 



1 200 

:-.':; OHO ^1 000 IV IK id 1 I OiHI II 000 ft 000 



Weight of Metal, in Pounds per Linear Foot of Bridge. 
Fig. 71. 



NICKEL STEEL FOR BRUXJES 



213 




'2 000 2 400 2 200 2 000 1800 1600 1400 1200 1 UUO 800 tWO 

Cost of Metal Erected, in Dollars per Linear Foot of Span. 
Fig. 72. 



214 NICKEL STEEL FOE BRIDGES 

APPENDIX A. PAPT I. 



COMPAKATIVE TESTS OF STEUCTUKAL NICKEL STEEL 
AND MEDIUM CARBON STEEL. 

Preliminary Work. — In November, 1903, this investigation had its 
inception, and, early in December, two parallel series of tests were 
made at the American Bridge Company's plant at Pencoyd, Pa. For 
this preliminary work, two steels of the following composition were 

Heat Heat 

No. 16 080 No. 17 005. 

Nickel 3.21 4.25 

Carbon 0.390 0.463 

Manganese 0.65 0.67 

Sulphur 0.015 0.014 

Phosphorus 0.011 0.019 

A full report of this examination was made in January, 1904, and 
the results were so promising that immediate steps were taken for a 
more elaborate series of tests upon a steel made especially for this pur- 
pose. Arrangements were made with the Carnegie Steel Company, 
in the same month, for two melts of nickel steel. 

After many delays it was finally decided to have the "shape" nickel 
steel rolled and tested before ordering the "eye-bar" nickel steel. Two 
melts of almost identical composition were made at the Homestead 
Works, Carnegie Steel Company ; the first. Heat No. 17 673, on No- 
vember 1st, and the second. Heat No. 17 749, on December 1st, 1904, 
each being of basic open-hearth steel. 

The following material was shipped to Pencoyd, Pa., for testing 
during December, 1904, and January, 1905: 

1 Universal plate, 12 by f in., length 17 ft. 

1 " " 12 by i in., " 17 ft. 7 in. 

1 " " 12 by I in., " 15 ft. 1 in. 

1 " " 12 by 1 in., " 17 ft. 

5 angles, 6 by 6 by f in., " 15 ft. 

1 " 8 by 8 by 1 in., " 15 ft. 

At the same time other material was shipped to Ambridge, Pa., 
for the fabrication of struts for compression tests and of eye-bars for 
"full-sized" tests. 

The medium-carbon steel plates for comparative tests were not re- 
ceived at Pencoyd vmtil December, 1905. They were as follows : 

1 Universal plate, 12 by f in., length, 20 ft. i in. 

1 " " 12 by i in., " 20 ft. 1 in. 

1 " " 12 by I in., " 19 ft. Hi in. 

1 « " 12 by 1 in., " 19 ft. 8J in. 

There were no tests of angles of carbon steel. 



NICKEL STEEL FOR BRIDGES 



215 



Composition of Steels. — Several analyses were made of both the 
nickel-steel and carbon-steel heats, as shown in Table 7. 

TABLE 7. — Analyses op Nickel and Carbon Steel. 



Nickel Steel. Heat No. 17 673. 





Nickel. 


Carbon. 


Manganese. 

0.70 

Grav. 0.80 
" 0.80 
^- 0.76 

Comb. 0.758 
0.771 


Sulphur. Phosphorus. 


Silicon. • 


Analyst. 


Desired . 


3.50 

3.66 
3.64 
3.66 

3.68 
3.87 
3.92 


0.38 


below 0.04 below (i.03 


below 0.04 




1 
2 
3 

4 
5 
6 


0.37 
0.37 
0.36 

J 0.408 ( 
1 0.407 ( 

0.471 
0.417 


0.025 1 0.010 
0.023 0.010 
0.084 1 0.010 

oen 1 ) 0.005 
"•0^0 I ( 0.005 

0.019 0.010 
0.021 0.0<% 


0.050 
0.047 
0.050 

Grav. 0.046 
Vol. 

0.039 
0.035 


Mill. 

(Shimer. 
■I Bethle- 
( hem. 
r Booth, 
J Garrett 
1 and 
I Blair. 
\ Orford 
1 Works. 



Nos. 1, 2, and 3 were ladle analyses, made at the mill during pouring, in the order 
named; No. 1 between the first and second ingots. No. 2 between the fourth and fifth 
ingots. No. 3 between the ninth and tenth ingots. The carbon was probably determined 
by gravimetric analyses. 

No. 4 was made from drillings taken from test ingot No 2. 

Nos. 5 and ti were made from duplicate drillings taken from %-in. plate. 



Nickel Steel. Heat No. 17 749. 



3.50 10.36 color. 
3.50 0.39 comb. 
3.50 0.38 color. 
3.52 0.356 comb. 



0.82 
0.78 
0.81 
0.763 



0.027 
0.033 
0.028 
0.030 



n.oii 

0.015 
0.011 
0.012 



0.060 
0.060 
0.048 
0.058 



Mill. 



Shimer. 



Nos. 1, 2, and 3 were ladle analyses, made at the mill during pouring, in the order 
named; No. 1 between the second and third ingots. No. 2 between the tenth and eleventh 
ingots. No. 3 between the tweltth and thirteenth ingots. 

No. 4 was made from drillings taken from test ingot No. 2. 



Carbon Steel. Heat No. 33 342. 
















f Booth, 
J Garrett 


1 




0.253 


0.546 


0.025 


0.014 


0.016 


2 




0.198 


0.55 


0.025 


0.011 


0.012 


1 Blair. 
S Orford 
■( Works. 


Nos. 1 and 2 were mad( 


J from duplicate drillings taken from %-i 


n. plate. 




Carbon Steel. Heat No. 41 520. 


1 


0.19 


0.54 


0.0:i3 


0.022 




Mill. 
( Booth. 


2 




0.287 


0.563 


0.025 


0.025 


0.020 


i Garrett 
' and 
I Blair. 


3 • 




0.234 


0.62 


0.024 


0.023 


0.035 


) Orford 
■) Works. 



Nos. 2 and 3 were made from duplicate drillings taken from ^-in. plate. 



216 NICKEL STEEL FOR BRIDGES 

For use in the fabrication of struts and for the 12-in. universal 
plates, tested for comparison with the nickel steel, a medium carbon 
steel of good quality was required. 

All the 12-in. plates for specimen tests were rolled from Nickel- 
Steel Heat No. 17 673 and Carbon-Steel Heat No. 33 342. The angles 
for specimen tests and the greater part of the material for the struts 
were rolled from Nickel-Steel Heat No. 17 749 and Carbon-Steel Heat 
No. 41 520. 

A comparison of the analyses of the steels shows that the material 
obtained was as close to the desired composition as it was practicable 
for the mill to make. The tests made on the plates did not reveal any 
lack of homogeneity. 

The surface of the nickel-steel plates was smooth and free from 
scale, much more so than that of the carbon-steel plates. The surface 
of the nickel-steel angles was not so good, as in one or two instances 
the fin from the edge and some scale had been rolled in. 

All the material shipped to Pencoyd and Ambridge was stored un- 
protected from the weather for several months, the best means for 
identification later, being the cleaner and smoother surface of the 
nickel steel. 

Tests of 12-in. Universal Eolled Plates and Angles. 

Tensile Tests — Plain Specimens. — The test given the greatest at- 
tention in mill practice everywhere is the tensile test. In the absence 
of inspection, or when, for any reason, knowledge of the physical prop- 
erties of material is desired by the mill, a tensile test only is made. 
At many mills no other test is made by the inspector for his client. 
For these reasons, a large number of specimens were prepared from 
this structural nickel steel and from the structural carbon steel tested 
for comparative purposes. 

Because of the uncertainty attached to the usual method of deter- 
mining the "yield point" or so-called "elastic limit," arrangements 
were made for two series of tests, one on parallel-sided pieces and one 
on pieces having edges machined parallel for a distance of 94 in., 
dnly, with fillet widening pieces at the ends, and known as the standard 
adopted by the American Association of Steel Manufacturers (abbre- 
viated here into A. A. S. M.). The length of all specimens was 18 in., 
and the width, except for those cut from the 1-in. material, was 1^ in. 
The width of these was reduced to 1^ in. because it was expected that 
a 150 000-lb. machine would be used to make the tests, and it was 
necessary to keep the ultimate strength well within this limit. 

The "parallel-sided" pieces were to be tested according to mill prac- 
tice at a representative mill. At the suggestion of Mr. C. L. Huston, of 
the Lukens Iron and Steel Company, Coatcsville, Pa., half these pieces 
were tested at this mill, and the others at the Pencoyd Iron Works, 



NICKEL STEEL TOR BRIDGES 217 

Pencoyd, Pa. Mill practice was not followed exactly, because it was 
not thought possible to get even a close approximation for the "yield 
point" of the nickel steel, with the high speeds in every-day use. The 
only change, however, was in the reduction of the speed of breaking. 

The A. A. S. M, pieces were to be tested in a machine equipped 
with an autographic attachment ; and a 200 000-lb. Olsen machine, at 
Drexel Institute, Philadelphia, was used, through the kindness of Mr. 
Earl B. Smith, Instructor in Mechanical Engineering. All the ma- 
chine work was done at the Pencoyd Iron Works. 

The tensile tests in the original lay-out were located at the ends of 
the plates and angles, the A. A. S. M. pieces at the edge, and the 
parallel-sided pieces inside and adjacent. As it was desirable to re- 
peat certain tests to bring out more definitely points of difference be- 
tween the two steels, additional pieces had to be selected from avail- 
able material. The tensile strength of plates varies with the location 
of the specimen in the plate and of the slab in the ingot, so this forced 
selection of additional pieces added necessary information on these 
points. All the material was rolled from slabs; hence these variations 
in physical properties due to location are not so pronounced as in ma- 
terial rolled directly from the ingot. 

American Association of Steel Manufacturers' Specimens. — The 
results of these tests are given in detail in Tables 30 and 31, and a com- 
parison with the results of tests on parallel-sided specimens may be 
found in that part of this appendix filed for reference in the Library 
of the Society. 

The speed of machine or movement of head for the first eleven tests 
was 1 in. in 6 min., but, while this gave good results for the carbon 
steel, it was too rapid for the nickel steel, and the yield point was lost. 
The weight could not be made to travel along the beam with suffi- 
cient rapidity to keep the beam balanced, and, consequently, the pencil 
recorded a faulty curve. A speed of 1 in. in 6 min. was desired in 
order to economize time, the speed, 1 in. in 20 min., adopted for the 
remaining pieces, requiring i hour for each test. In these remaining 
tests, up to a point just below the yield point, a speed of 1 in. in 6 
min. was used, and then a speed of 1 in. in 20 min. until the material 
had well passed this point, when 1 in. in 6 min. was again used. In 
several tests, when the maximum strength had been developed, a still 
faster speed (1 in. in 2 min.) was used for breaking the specimen. 
The speed at breaking was found to affect the tensile strength ma- 
terially. 

Yield Point. — The term "yield point" has been adopted herein in 
place of "elastic limit" because it more nearly represents what was 
sought, namely, the point at which the material would no longer re- 
turn to its original condition upon the removal of the load. It is 
shown graphically in the autographic diagram by the departure of the 
record from a straight line. 



318 



NICKEL STEEL FOR BRIDGES 



The determination of the true elastic limit was not possible with 
the apparatus at hand. 

The loss of time caused by the use of an autographic attachment 
bars the adoption of this method in mill practice, yet it was desired 
that the yield point reported in these tests should be a point easily 
obtainable in practice by a skillful operator, and not one depending 
upon time-consuming and expensive devices for its exact location. It 
was assumed to be the "least permanent set" visible in a specimen by 
the aid of a pair of finely pointed dividers. The point in the curves 
corresponding to this "least permanent set" was obtained after some 
experimenting, as shown in Table 8. 

TABLE 8.— Beam Readings. 



Mark. 


1 

! Drop of beam, 


Yield point, 


Load 


Permanent set in 6 in. 


in pounds. 


in pounds. 


released at: 


by dividers: 


CT8L 150 


38 COO 


35 200 


36 600 lb. 


0.01 in. 


•' 193 


36 600 


34 800 


36 000 " 


0.02 '• 


" 198 


39 400 


33 800 


33 200 '• 


0.01 " 


" 153 


34 200 


33 000 


33 600 " 


0.015 '• 


TSL 3 




38 000 


38 000 " 


0.005 " 


" 40 




39 300 


39 000 '■ 


0.015 " 


" 43 




40 600 


40 000 " 


0.01 " 


" 44 




38 800 


39 000 " 


0.015 " 


" 58 




46 500 


48 000 " 


0.08 •' 


" 70 




47 100 


46 900 " 


0.005 " 


'• 89 




47 500 


48 000 " 


0.02 '^ 


" 109 




70 500 


69 300 " 
75 000 '• 


0.00 " 
Between these readings 
a rapid elongation of 
specimen took place. 


" 150 




77 400 


77 800 " 


0.035 in. 


" 154 




7t! 600 


77 400 " 


Appreciable. 


" 199 




73 000 


72 000 " 
76 000 " 


Not appreciable. 
0.01 in. 




i 


69 200 


64 400 " 


Not appreciable. 


ATSL 58 


\ :::::::; 




70 600 " 
75 000 " 


Not appreciable. 
0.02 in. 



In the tests given in Table 8 the load was removed at about the 
yield point, the permanent set, if any, was measured, and the move- 
ment of the pencil noted. If this traced a new line parallel to the 
original curve, there was visible evidence of a permanent set and of 
the amount. To check this, careful measurements were made by 
dividers. It was assumed that, for all practical purposes, a set of 0.01 
in. in 8 in. was the "least visible," and the point in the curve corre- 
sponding to this was obtained graphically at the intersection of a line 
0.01 in. away from, and parallel to, the straight portion of the record. 
The stretch shown by the curve is the stretch in 8 in. It was found 
that, even when the record had departed from a straight line, if ap- 
proximately this distance of 0.01 in. had not been exceeded, on releas- 
ing the load the pencil would retrace the original line, and no elonga- 
tion could be detected by the dividers. 



NICKEL STEEL FOR BRIDGES 219 

At times the mechanism of the recorder did not work properly, 
but usually these vagaries were easily corrected. The movements of 
the weight and of the pencil were constantly checked, one with the 
other. The yield point thus determined is slightly lower than the drop 
of beam. In tests of nickel steel, of this content of nickel and carbon, 
the beam does not drop at the point where the stretch continues with- 
out an Increasing load, because there is no such point. The auto- 
graphic curves show this; for it will be seen that in no curve of nickel 
steel is there a bend at the yield point of even approximately 90 de- 
grees. A "drop" will occur if the weight is run out uniformly, for the 
rate of stretch changes, but it will be almost invariably at too great a 
load as registered by the weight on the beam, for in mill practice the 
speed is so great that the beam is never balanced until after the so- 
called elastic limit or yield point is reached. In carbon steel an op- 
portunity is given, by the stretching of the steel, for the weight to 
"catch up"; in nickel steel there is none. This subject will be men- 
tioned later. 

The elongations in 2, 4 and 6 in. are given; although their value is 
not great, they show that considerable stretching takes place over the 
entire length of the specimen. 

The "area of fractured section" was measured at points midway be- 
tween edges at Pencoyd and Drexel. This method is followed in most 
mills, but it gives slightly higher percentages of reduction than actually 
exist. At Lukens, edge measurements Were taken, that being the prac- 
tice at that mill. The dimensions of this section, and also of the orig- 
inal section, are given to the nearest 0.005 in., this being near enough 
for all practical purposes. At Pencoyd and Drexel the elongations 
were measured along the edge of the specimen; at Lukens, along the 
middle. This difference affects slightly the comparison of results. The 
elastic ratio is the ratio of yield point or drop of beam to the ultimate 
strength. 

In Table 31 both drop of beam and yield point, for specimens 
broken at Drexel Institute, are given. The "stretch at the drop of 
beam," in Table 31, has no special significance, except to show why 
mill methods will give fair results for testing carbon steel and not 
for nickel steel, in which no such stretch occurs. It was measured 
from the autographic records. The drop of beam was taken from the 
cards as that load at which the curve first became horizontal. Actually, 
no drop of beam occurred, as the machine was equipped with a device 
for controlling automatically the movement of the weight; and, when 
adjusted properly at the slow speed used, the beam was balancing 
throughout. 

Parallel- Sided Specimens. — The tests on parallel-sided specimens 
are interesting, as showing how close an approximation to the true 
yield point may be obtained by the drop of the beam, by expert 
operators of mill-testing machines. 



220 



NICKEL STEEL FOR BRIDGES 



A 200 000-lb. Olsen machine was used at each mill. At Pencoyd 
the weight was run out on the beam by an electric motor, excited by 
contacts at the end of the beam, and, at Lukens, by a hand-wheel. 
Operating by hand is shown to be much more sensitive, the reason 
being easily seen by one witnessing a test. When testing the nickel 
steel, the Pencoyd operator had to know in advance about where to ex- 
pect the drop. Even with this information, the results are largely 
guesswork. The Lukens operator did not need this information, and, 
while the beam readings are a little high, for the reasons explained, 
the results are in fair agreement with those obtained with the auto- 
graphic recorder, and the time for making the tests is brought within 
the possibilities of mill practice. The speeds used at each mill for 
nickel steel were reduced below those for carbon steels. 

Examination of Eesults — Nickel Steel — Tables 30 and 32. 

A fair degree of uniformity exists throughout. 

Yield Point. — Edge specimens have a higher yield point than cor- 
responding interior pieces. Specimen tests made at Lukens agree 
closely with those made at Drexel, but those made at Pencoyd are gen- 
erally higher. The yield point becomes lower as the thickness of the 
material increases. 



Material 


%in. plate. 
20 min. 


)^-ln. plate. 
20 min. 


^-in. plate. 
20 min. 


1-in. plate. 
20 min. 


•M-in. angle. 
20 min. 


1-in. angle. 


Speed, 1 in. in: 


20 min. 


Edge 


70100 
67100 


62i66 


62 100 


58 600 
58 500 


62 300 


54 2(K) 











Tensile Strength. — Edge specimens have a tensile strength about 
3 000 lb. higher than corresponding interior pieces. All specimens cut 
from the same thickness agree closely when variations of speed and 
location are considered. The tensile strength becomes less as the 
thickness increases, also as the speed of breaking decreases. 

TABLE 9. 



Material... 


%-iN. Plate 


i^-iN. Plate. 


%-iN. Plate. 


Speed, 1 in. in: 


a 

S 
to 


2 


40-10 sec. 


d 

a 

CD 


3-2 min. 


40-10 sec. 


a 

s 

to 


d 
B 


o 

o 

5 


Kdge 


116 600 
11;^ 900 


116 300 
112 600 


lV(i lOO 


110 600 


113 800 
107 500 


112 500 


100 8(X) 


108 400 
108100 




Interior 


111 200 


Material 


1-IN. Plate. 


•M-iN. Angle. 


1-IN. Angle. 


Edge 

Interior 


104 600 


107 300 
103 500 


105 Vob 


103 000 




104 400 




97 900 


lod sdo 



NICKEL STEEL FOR BRIDGES 



221 



1-in. and |-in. plates were rolled from slabs 14^ by 9^ in. 

i-in. " i-in. " " " " " m by 8i in. 

8 by 8 by 1-in. angles " " " " 11 i by 8 in. 

6 by 6 by |-in. " " " " " 8 by 9 in. 

The slabs were rolled from ingots 25 by 30 in. 

The relative amounts of work done on the slab in rolling are shown 
in Table 10. 

TABLE 10, — KELATnE Amounts of Work Done in Kolling. 



l2-in. plates, 
thickness. 


Area of slab. 


Area of shape. 


Percentatje of 
reduction. 


1 in. 

Angles. 
8 X 8 X lin. 
6 X 6 X M in. 


134 sq. in. 
134 •' 
120 " 
120 '■ 

92 " 

72 " 


12 sq. in. 
9 
6 

4^ '• 

8^ •' 


91.0 
S3.3 
05.0 
96.3 

as.i 

88.2 



The differences in tensile strength due to thickness alone result 
probably from a combination of two varying conditions : first, the 
chilling during rolling; and, second, the amount of reduction of the 
slab. Differences in other physical properties necessarily arise from 
these same conditions. 

Elastic Ratio. — The elastic ratio is seen to decrease uniformly as 
the thickness increases, varying from 60.1 for the f-in. material to 
54.5 for the 1-in. material. 

Elongation and Reduction of Area. — The elongation and reduction 
of area increase slightly as the thickness of the material increases, 
but the variation lacks uniformity. 



TABLE 11. — Elongation and Eeduction of Area. 




Material 


%-iN. Plate. 


1^-iN. Plate. 


%-m. Plate. 


Speed, 1 in. in: 


d 

s 


p 
B 


i 

03 
O 

i 


p' 
S 

OS 


3-2 min. 


o 
o 


50 


3-2 min. 
40-10 sec. 


„, .. ( Edee ' 


15.7 
16.5 
46 4 


"isle" 


16.5 


16.4 
17.1 
46.0 
50.0 


"ih'.h" 

■■47!3" 


17.5 

■48!i' 


18.1 


Elongation: \^J^^[^- ^r,/^' 
Reduction! (Edge.... 


16.5 17.8 
47 4 


of Area: )' "i Interior.! 45.6 


48.9 


45.9 j 


43.6 40.7 


Material 


] 
1-IN. Plate 5i-iN. Asgle. 


1-n 


«. Angle. 


„, . . \ Edge 




20.0 
20.5 
47.4 
49.8 


18 7 








21.3 


Elongation: , j^tlrior . 


19.8 


21.2 
"48!5"' 






17.9 




18.4 


Reduction | * Edge 


42.0 






48.5 


of Area: f "( Interior . 


49.2 




35.2 




41.0 



















222 



NICKEL STEEL FOR BRIDGES 



Speed of Machine.- — The tensile strength is raised by an increase 
in speed; the effect on elongation and reduction of area of fracture is 
not marked; the tendency, however, is to reduce the reduction of area. 

Examination of Eesults — Carbon Steel — Tables 31 and 33. 

From only one thickness of material, 1-in. plate, were specimens 
cut from both edge and interior, hence remarks as to the effect of 
location are limited to this one example. 

Yield Point. — Edge specimens have a yield point slightly higher 
than the corresponding interior pieces. 

Drop of Beam. — 

The average drop of beam for all Drexel tests is 34 100 lb. 
a a u « u u x^ukens " " 32 800 " 

" a a a u u u Pgncoyd " " 40 600 " 

The speed of machine differed in each case, being 1 in. in 6 min. 
at Drexel, 1 in. in 3 min. at Lukens, and 1 in. in 15 sec. at Pencoyd, 
yet the close agreement between the two former and the great differ- 
ence of the Pencoyd readings show how little dependence can be put 
upon results obtained with the machine running at a high speed. 

The average of all the Drexel tests for the yield point is 32 800, 
which is 1 300 lb. lower than the corresponding drop of beam. 

There is a close uniformity between the results obtained at the 
same place. 



' 


plate. 


plate. 


•M-in. 
plate. 


lin. 
plate. 






Yield point at Drexel 

Droyj ot beam at Drexel . . 
Drop of beam at Lukeiis. 
Driip of beam al Pencoyd 


38 600 

40 800 
38 000 
46 700 


36 900 

37 300 
33 900 
40 800 


33 400 

34 100 
30 900 
36 400 


27 800 
30 300 
38 500 
38 500 


(26 100)* 
(27 000)* 


Edge. 
Edge. 
Interior. 
Interior. 










* Interior specimens. 

These figures show how regular is the decrease in the yield point 
as the thickness of the material increases. The same conditions prob- 
ably exist as influenced the physical properties of the nickel steel. 
The edge specimens cut from the 1-in. plate have a higher yield point 
than the corresponding interior ones. 

Tensile Strength. — Edge specimens have a tensile strength about 
:> 000 lb. higher than the corresponding interior pieces. 



Material... 


a 
6 


N. Plate. 


'/6-iN. Plate. 


•$4-iN. Plate. 


1-IN. Plate. 


Speed, 1 in. iu:. 


i 


en 

o 

60 166 


a 


.3 
g 

05 


c 


d 
S 


d 
S 

OS 


o 

2 


6 min. 
3-2 min. 


o 
in 






63 700 
62400 


63IH00 






62200 






61900 








62 000 


65 366 


61100 


62 700 


59000 58300 


61600 










; 





NICKEL STEEL FOR BRIDGES 



223 



The differences due to thickness are evident, but they are small, 
and those due to speed of breaking and of location in plate are well 
marked. 

Elastic Ratio. — The elastic ratio decreases as the thickness of ma- 
terial increases, varying from 60.8 for the i-in. material to 45.0 for 
the 1-in. material. 

Elongation and Reduction of Area. — The elongation seems to in- 
crease slightly as the thickness increases, but the effect is small, and 
it is not noticeable at all on the reduction of area. An increase in the 
speed of breaking causes a decrease in the reduction of area, although 
this is not well marked, and the effect on the elongation is not notice- 
able. 



Material, 



i-iN. Plate. 



Speed, 1 in. in:. 



El'^'igation IKior! 



Reduction of area. I ^jftlrior," 



28.0 
28.5 



56.2 
59.5 



56.1 



}^-iN. Plate, %-in. Plate, 



27.4 
56.5 



30.5 
59 .'2 



....82.2 
28.0'.... 



55.3 



58.0 



c ■ 2 



1-IN-. Plate. 



31 9 
32.5 33.7 33.5 31 ;0 34!4 



... ....I.... 56.7 .... 

59.155.9 60.4 61.4 57.2 



Comparison of Nickel Steel and Carbon Steel. 
The varying conditions affecting the physical properties of steel 
seem to have no greater effect on one metal than on the other, except 
upon the tensile strength. The greater relative reduction in strength 
of nickel steel with increasing thickness of material causes the elastic 
ratio for the thicker metal to be higher for the nickel steel than for 
the carbon steel, as shown by the comparisons in Table 12. 

TABLE 12.— Elastic Eatio. 





%-in. plate. 


^-tn. plate. 


%-m. plate. 


1-in. plate. 


Tested 
at: 


Speed, 
1 in. in: 


Location. 


Nickel steel 


60.1 


58.0 


57.8 


*54.5 


Drexel. 


6min. 


Edge. 


Carbon steel 


*60.8 


57.8 


53.7 


*45.0 


" 


6min. 


Edge. 


Nickel steel 


60.0 


.56.3 


62.6 


57.0 


Lukens 


3 min. 


Interior. 


" 


58.1 


55.1* 


56.1 




" 


10 sec. 




Carbon steel 


*64.4 


.58.5 


54.8 


*49.0 


Drexel. (Iniin. 


Edge. 


;. 


60.7 


55.2 


51.1 


48.7 


Lukens 


3 min. 


Interior. 




.58.6 


51.4 


48.7 


46.1 




10 see. 





* Speed, 1 in. in 2 min. 

Nickel steel is not as ductile as carbon steel, yet it is not brittle 
in any sense. The yield point, or, as it is usually called, the elastic 
limit, may be safely taken as a minimum at 60 000 lb., as compared 
with the 30 000 lb. usually specified for carbon steel. In tests of the 
1-in. material, a lower value was obtained, but the speed was slower 



224 NICKEL STEEL FOR BRIDGES 

than in mill practice; and the carbon steel, tested similarly, gave also 
a lower value than 30 000 lb. 

The ultimate strength varies with the thickness, speed, and the 
location of the specimen, but, as compared with the results of carbon 
steel, a value between 100 000 and 120 000 lb. may be obtained, and 
even 110 000 lb. as a minimum, if the conditions mentioned are not 
restricted — and it is not usual to restrict them. 

An elongation of 15% in 8 in., and a reduction of area of 40% 
may be safely taken as the minimum. In almost every instance the 
fractures were silky, uniform, and free from laminations, pipe, etc. 
The only variation was in four of the plate tests and in the |-in. 
angle, these fractures being partly fine crystalline or granular. 

The usual high elastic ratio for nickel steel is not found in these 
tests, probably because the speed of machine was much slower than is 
generally used, and the determination of the yield point was more 
accurate. As already stated, the methods in vogue at most mills 
would give a very high but inaccurate value for the yield point or 
elastic limit of nickel steel. 

The tests were made at Drexel in February and March, at Pen- 
coyd in March, and at Lukens in April, 1906. 

Tensile Tests of Punched, Reamed^ and Punched-Riveted 
Specimens. 

These tensile tests were arranged, as far as possible, to show 
relatively what changes are effected in the physical properties of 
nickel steel and carbon steel by the shop operations of punching holes 
full size, of sub-punching and reaming to full size, and of riveting 
punched work. The full-sized holes were \^ in. in diameter, the sub- 
punched holes, ]-l in. in diameter; ^ in. of material, therefore, was 
reamed away. 

It is hardly necessary to repeat here that the usually accepted 
opinions are that shearing the metal, as in punching, forms numerous 
incipient cracks radiating from the edge, and also that the material 
around the hole is hardened by the pressure necessary to force the 
punching out. It is believed that if these cracks and this hardened 
ring are cut away, all injury done to the material is removed. The 
opinions as to the effect of riveting are, perhaps, not so positive, be- 
cause the subject has not been investigated experimentally ; it was 
only introduced into this series after other tests had been made. The 
object was to determine the annealing effect, if any, on the material 
due to the driving of the hot rivet. 

Test specimens, 3 in. wide and 18 in. long, were cut from the 
i-in. and the |-in. plates. The "punched" and the "reamed" pieces 
were laid out side by side in every instance, in order to give results 
as directly comparable as possible; the "punched-riveted" pieces were 
cut from available surplus material. Through an error, one piece of 



NICKEL STEEL I'OK BRIDGES 



225 



carbon steel, which it was intended to punch full size, was sub- 
punched and reamed, and vice versa. The edges of all holes and plates 
were rounded slightly by filing to remove burrs. 

Pieces TPPL 13 and 33, TAPPL 46, 140, and 142 were cut from 
the edge of the plate, but this does not seem to have influenced the 
results as in the tensile tests of plain specimens, probably because of 
the §:reater width of 3 in. The punching, reaming, and riveting were 
done in the bridge shop at Pencoyd in the customary way. The diam- 
eter of the die used in the punching was ^^ in. larger than that of 
the punch. To prevent injury to the material in cutting off the heads 
of rivets, the latter were driven with a plate, 3 by xij by 9 in., on each 
side of the specimen. These plates served another purpose, not in- 
tended, namely, the lessening of the amount of compression in the 
material under the head of the rivet, caused by the pressure of the 
cup of the riveting machine. The pressure on the rivet in driving was 
probably about 30 tons per sq. in., and the annealing effect, if any, 
was entirely destroyed by the compression around the hole. The metal 
under the head was bright, and the calipering showed a reduction of 
thickness averaging as follows : 

Carbon steel. Nickel steel. 

f 0.00:5 in. 0.003 in. 

f 0.007^ " 0.003 " 

All the testing was done on the testing machine at Drexel Insti- 
tute, and the autographic recorder was used to plot the curve of each 
test; and from this record the elastic limit, yield point, and ultimate 
strength were determined. 

Yield Point. — As in the tensile tests of plain specimens, at about 
the yield point, the load was removed from several specimens and the 
permanent set carefully measured by dividers and checked with the 
card record, the results shown in Table 13 being obtained : 





TABLE 13.— Beam Readings. 


Mark. 


Yield point, 
in pounds. 


Load released at: 
Pounds. 


Permanent set in 6 in., 

by dividers: 

Inches. 


CTAPPL 46 


29 500 


21 000 


None. 






.30 200 


Widened line. 


140 


66 500 


66 000 


None. 






68 000 


Widened line. 






72 000 


0.05 


143 


68 000 


70 000 


0.02 


TAPPL 46 


62 000 


46 700 


None. 






63 000 


None. 






73 000 


0.03 


140 


128 000 


112000 


Widened line. 


142 


lao 000 


120 000 


0.01 


Heat No. 16 080: 








TPPL 3/8 


53 200 


54 000 


0.01 scant. 


TPPL 3/4 


100 000 


06 000 


Widened line. 






104 600 


0.02 scant. 


TPRL 3/8 


52 000 


52 000 


Widened line. 






56 000 


0.02 scant. 


TPRL 3/4 


98 800 


90 000 


Widened line. 






102 800 


0.015 



226 NICKEL STEEL FOR BRIDGES 

The yield point, as before, was obtained from the curve at the in- 
tersection of a line 0.01 in. away from, and parallel to, the straight 
portion of the record. While there is a fair agreement between the 
results obtained by the two methods described, it is not as close as in 
the tensile tests on plain specimens, and an examination of the curves 
will show the reason. The deflection from a straight line is so gradual 
and, in those of the nickel steel, so slight, that it is impossible to ob- 
tain, by any practical means, the exact point where a permanent set 
first takes place. Yet the loads recorded in these tables are closely 
approximate (with two or three exceptions mentioned later), so that, 
for comparative purposes, the data obtained are of sufficient value. 

In all the curves of the carbon steel there is a decided change of 
direction, corresponding to the stretch occurring at the yield point 
for the plain specimens. In these tests this change varies in form and 
position. In the tests of the f-in. material it occurs just before the 
maximum strength is developed; and, in those of the |-in. material, 
it is found close to the yield point. There is a sufficiently marked 
change in the curve at what has been taken as the yield point to sub- 
stantiate the tabulated figures. In the curves for tests from Heat No. 
16 080, similar changes occur, but not in the other nickel-steel tests 
— the former material is a softer grade of steel. The |-in. material 
is for all steels softer than the -|-in. material; the variations described 
may be assumed, therefore, as being conditional upon the grade of 
steel. 

All these tests were made in a similar manner, a speed of 1 in. in 
20 min. being used just before and until the yield point was well 
passed, and a speed of 1 in. in 6 min. at the beginning and at the 
instant of breaking. 

Tables 34, 35, and 36 show the results of each test in detail. 

1. Curve CTPBL No. 7 differs from the other three in that it 

has a higher yield point and is straighter within this 
point. The autographic attachment did not work prop- 
erly. 

2. Curve TPRL No. 7 is slightly straighter within the yield 

point than the other three curves, but it is not abnormal. 

3. Curve TPPL No. Ill differs for the same reason — the curve 

is flatter. 

4. Curves for CTAPPL are all irregular; the autographic at- 

tachment did not work properly in describing any of 
these curves, but the yield point is well marked, so that 
the results are fairly accurate. 

5. Curve TAPPL No. 141 differs from the others for the same 

reason. These curves are all so flat that it is practically 
impossible to obtain the yield point accurately. 
The other curves are all normal. 



NICKEL STEEL FOR BRIDGES 



227 



Examination of Results.— Ti\h\es 34, 35, and 36. 

In the comparisons in Tables 14 and 15 it must be borne in mind 
that these tests were made on specimens 3 in. wide, while the plain 
specimens were but 1^ in. wide. It is well known, also, that short 
grooved specimens give a higher ultimate strength than the usual 
18-in. specimen. It is not probable, therefore, that all the effects de- 
scribed are the direct result of alterations in the structure of the ma- 
terial about the hole. 

TABLE 14. — Percentages Based on Results of Tests of Beamed 

Specimens. 





Physical property. 


%-iN. Material. 


%-iN. Material. 


steel. 


Reamed. 


Punched. 


Punched- 
riveted. 


Reamed. 


Punched. 


Punched- 
riveted. 


Nickel. . 




95 
102 
101 
103 

04 

lUO 


102 
105 
91 
93 
112 
113 


113 

107 

94 

93 

121 

115 


114 
113 
104 
107 
110 
104 


Ill 
124 
89 
94 
125 
132 


130 


Carbon. 
Nickel.. 
Carbon. 
Nickel.. 


Ultimate strength 

Elastic ratio 


128 
82 
81 

157 


Carbon. 




158 









TABLE 15. — Percentages Based on Results of Tests of Plain 

Specimens. 





Physical property. 


%-iy. Material. 


%-lN. Materiajl. 


steel. 


Punched. 


Punched- 
riveted. 


Punched. 


PuDched- 
riveted. 


Nickel 


Yield point 


108 
103 
91 
91 
120 
114 
62 
77 
51 
75 


120 
105 
93 
91 
130 
115 
91 
80 
57 
64 


97 
111 
85 
88 
114 
126 
50 
39 
39 
28 


113 


Carbon 




114 


Nickel 


Ultimate strength 


79 


Carbon 




75 


Nickel 


Elastic ratio 


143 


Carbon 




152 


Nickel 


Elongation iu 4 in 


46 


Carbon 




14 


Nickel 


Reduction of area 


22 


Carbon 




3 









Punching, suh-punching-and-reaming, and riveting, in every in- 
stance, alter the |-in. material more than they do the §-in. material, 
in both nickel and carbon steels. 

Punching ^^-Inch Hole. — Punching raises the yield point and 
lowers the ultimate strength. The yield point of the nickel steel is 
affected to a smaller extent than that of the carbon steel, and the ulti- 
mate strength more, but the difference is not great. 

Reaming — From W-Inch to tH -Inch. — The process of sub-punching- 
and-reaming raises both the yield point (with but one exception, 



228 



NICKEL STEEL FOR BRIDGES 



namely, that of the i-in. nickel steel) and the ultimate strength of 
both steels, and in almost the same proportion. The effect is less than 
that of punching full size, probably because the punched hole is 
smaller, and the injury, therefore, less, and because the reaming re- 
moves the greater part of the affected area. Nickel steel, on the 
whole, is affected to a smaller extent than carbon steel, but the differ- 
ence is not uniform, and is small. The elongation and reduction of 
area cannot be compared directly with those of the plain specimens be- 
cause of the marked difference in the shape and character of the 
specimen. Comparing the results of punching with those of sub- 
punching-and-reaming, however, shows that the ductility of the ma- 
terial is injuriously affected by punching without subsequent ream- 
ing. The difference between the effects of punching and reaming 
upon the f-in. nickel steel is greater than it is upon the corresponding 
carbon steel, but it is less upon the |-in. material. 

Riveting. — Riveting raises still further the yield point, and lowers 
the ultimate strength. The elongation and reduction of area are also 
smaller than for the reamed specimens. Upon the g-in. material, how- 
ever, riveting seems to be beneficial, but upon the |-in. material it 
adds to the injury resulting from punching. Nickel steel is affected 
much less by riveting than carbon steel, probably because its resistance 
to compression is greater. 

The |-in. material being a softer steel than the f-in., it was to be 
expected that the effect of riveting upon it would be greater. 

In comparison with the carbon steel, the nickel steel in these special 
tests shows up very favorably, on the whole. 

TABLE 16. 



Kind of specimen. 



^-iN. Material. 
Area = 1.125 sq. in. 



Actual Actual 
yield ultimate 
point, streugth. 



Elastic 
i-atio. 



%-iN. Material. 
Area = 2.25 sq. in. 



Actual 
yield 
point. 



Actual 
ultimate 
strength . 



Elastic 
ratio. 



Nickel Steel. 


Plain specimens 


76 000 
51000 
55 000 
61000 


128 000 
91000 
82 000 
85 000 


59 
56 
67 

72 


140 000 
110 000 
]0() 000 
124 000 


243 000 
174 000 
147 000 
137 000 


58 




03 




72 


Punched-riveted specimens. 


90 


Carbon Steel. 




42 000 

29 000 

30 000 
30 000 


70 000 
48 000 
43 000 
43 000 


CO 
60 
70 
70 


72 000 
57 000 
64 000 
67 000 


137 000 
102 (ICO 
90 000 
78 000 


53 




50 




71 


Punched-riveted specimens. 


86 



NICKEL STEEL FOR BRIDGES 



329 



The comparisons (in Table 16) of the actual yield point and ulti- 
mate strength of these specimens with those of the plain specimens, on 
the basis of a sectional area of 3 by f-in. and 3 by f-in. show that the 
loss of strength is less for the nickel steel in |-in. material than for 
the corresponding carbon steel, and greater for the |-in. material. 
The differences are small. 

TABLE 17. — Percentage of Loss, Based on Plain Specimens. 





Physical Property. 


%iN. Material. 


1 


. Material. 


SteeL 


'i 


'6 


•a . 

Ol-rt 


i 


«3 


"Sh 








.a 


^« 


.a 


MS 






g 


ti 


o-iS 


u 


oZ 






§ 


a 


? 


« 


a 


a <D 






1 

Oh 


<s 


£■6 


3 
Ph 


a > 
cut. 


Nickel 


Yield point 


.33 


28 


20 


21 


24 


11 


Carbon 


" 


31 
29 
81 


29 
36 
39 


29 
34 
39 


21 
28 
26 


11 
40 
34 


7 


Nickel 


Ultimate strength 


44 


Carbon 


43 



TABLE 18. — Comparison of Assumed Yield Point with Elastic 

Limit. 

(In Pounds per Square Inch.) 



■3 


Physical 


Nickel Steel. 


Carbon Steel. 








„• 


"3 . 




T3 


•a 


•^ . 


i 

a 


property. 


_a 
'3 


4) 




«5 


a 
'S 




.a 
o 












n 


^ 









s a> 








^ 


3 
&H 


£-^ 




P5 


^ 


3 > 
Ph'S 


%-in... 


i Elastic limit.... 


60 000 


49 000 


50 000 


55 000 


35 000 


28 000 


30 000 


27 000 


1 Yield point 


67 000 


64 000 


69 000 


77 000 


39 000 


40 000 


41 000 


41 000 


M-in. . . 


\ Elastic limit.... 


.53 000 


53 000 


51 000 


61 000c 


29 000 


20 000 


32 000" 


35 000b 


/ Yield point 


62 000 


71000 


69 000 


80 000 1 


33 000 


37 000 


42 000 


43 000 



a. No. 126 not included. 

h. Pencil of autographic recorder did not work properly; the figures, therefore, may 
not be accurate, 

e. Curves are so (hit that the figure may not be accurate. 

Table 18 shows that within the yield point there are changes in the 
material, as shown in the autographic curves. The departure of the 
record from a straight line is what is meant by elastic limit. 

The number of tests was too small, and the apparatus used was not 
sufficiently delicate to furnish much more than a suggestion that these 
operations, necessary to the fabrication of bridge members, weaken the 
perfect elasticity of the material. The ductility is also lessened. The 
method used for determining the yield point in these tests does not 
give such uniform or accurate results as in the tests of plain speci- 
mens. In the specimens, the ratio of stretch to the load increases very 



230 NICKEL STEEL FOR BRIDGES 

gradually in these tests, especially in the case of nickel steel, the rec- 
ords of which are almost straight. 

The exceptional results shown for the |-in. material may have been 
caused by the bending of the test pieces during punching and riveting. 
This treatment would have a tendency to raise the recorded value for 
the elastic limit, though actually it would be less. The results for the 
punched-riveted specimens of |-in. material may be inaccurate as 
noted. 

The fractures are interesting. The f-in. material, on the whole, 
was silky, though the nickel steel in the punched and riveted specimens 
was partly crystalline. 

The |-in. material of nickel steel was fine crystalline, as was also 
that of carbon steel, with the exception of the reamed specimens. The 
nickel steel, however, was much finer-grained, and showed more plainly 
lines like magnetic lines radiating from the hole. Under the skin the 
fracture was silky, and this was more marked in the nickel steel. The 
material near the hole was also silky, but this was more marked in the 
carbon steel. 

The break occurred, with two or three exceptions, simultaneously 
on both sides. The difference in any case was only a fraction of a 
second. Those with a crystalline fracture were always unexpected, 
with no drawing down of the breaking section. 

The tests of the punched and of the reamed specimens were made 
in March, and those of the riveted specimens in June, 1906. 

Bending Tests, Plain Specimens. 

Next to the tensile test, the test most frequently made is the bending 
of a specimen piece of steel until it is proved that the ductility of the 
material is sufficient to satisfy the specifications. This test is also an 
indication of the absence of high phosphorus. 

The bending may be done in various ways: On a U-shaped block, 
under a steam hammer, by being pushed through an opening in an 
anvil by a plunger, or, in the case of iron and soft steel, by being 
wrapped around a mandrel, one end of the specimen being held rigidly 
and the power being applied at the other end by an eccentric cam. 

This last method makes the most perfect bend, but for heavy sec- 
tions of medium-carbon steel and especially of nickel steel, it is not 
practicable. The method to be followed is not often specified, yet it is 
an important factor, as is the speed of machine in the tensile test or 
the taper of the drift-pin in the drifting test. 

These tests were made in the testing room of the Lukens Iron and 
Steel Company through the kindness of Mr. C. L. Huston, Vice-Presi- 
dent, and Mr. Howard Taggart, Engineer of Tests. 

The specimen was laid on an anvil over a rectangular opening, 
wider at the top than at the bottom. The piece was bent by being 



PLATE XV. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1103. 

WADDELL ON 

NICKEL STEEL FOR BRIDGES. 





NICKEL STEEL FOR BRIDGES 331 

forced through by a plunger with its end rounded to any desired radius. 
The opening may be made of any desired width within certain limits. 
The plunger is attached to the piston of a hydraulic cylinder, conse- 
quently the pressure is very uniform. The piece is only partly bent, 
however, and is then flattened by the same machine. 

For large, thick pieces a plunger 4^ in. in diameter is used, but 
ordinary structural material, up to 1 in, in thickness, is bent about a 
2-in. plunger or mandrel. 

The specimens for these tests were all 2 in. wide by 18 in. long, 
two being cut from the end of each plate, one edge of each specimen 
being the mill edge, finished in a universal mill, the other edge being 
planed. Through the carelessness of the machinist, several of the pieces 
were planed on both edges. One piece of nickel steel, BL 186, cut from 
the 1-in. plate, was lost, and a piece cut from the interior, marked BL 
196, was substituted. All edges were smoothed by filing to remove burrs. 

One-half of the total number of test pieces were first bent by the 
2-in. mandrel, and afterward flattened until they cracked, or to 180° flat. 
This was done in order to find out how much bending the material 
would stand. All the carbon-steel specimens bent 180° flat, or until the 
capacity of the machine was reached, with a slight crack appearing in 
one piece only. The nickel steel also bent 180°, but cracked before 
the inner surfaces came in contact. With one exception, the nickel 
steel bent with a maximum radius of bend, such as would be formed 
about a mandrel having a diameter equal to 1.7 times the thickness of 
the material. One piece of angle material, ABL 61, was bent about the 
4J-in. mandrel, but the bend was not so good. 

The results are shown by the photograph, Fig. 1, Plate XV. 

It seemed from these tests that the material would stand bending 
about a mandrel of a diameter equal to twice the thickness of the mate- 
rial without cracking; hence four of these mandrels were made, one 
for each thickness of material. If the material had bent to the diame- 
ters intended, probably the results would have been as expected, but 
the bending was too localized and an acute angle was formed, so that 
in no instance was the radius of the bend the same as that of the end 
of the mandrel. Cracks, therefore, were developed, the amount of 
bending being practically the same as before. One piece, BL 85, from 
the i-in. .plate, broke, and one-half was evidently thrown on the scrap 
pile and destroyed. The |-in. and 1-in. material did not show any 
cracks. In order to prove positively that the f and i-in. plate material 
and the 1-in. angle material would stand a bending of 180° without 
cracking, with a diameter of inner surface of bend of 2t (t = thick- 
ness of material), the broken ends of some tensile test pieces were tried. 
This is considered a severe test, because of the straining, beyond the 
elastic limit, of all parts of the piece in tension. Many of the pieces 
were badly cut by the wedges. 



232 NICKEL STEEL FOR BRIDGES 

The bending was done at Pencoyd under a small steam hammer. 
The load, however, was not applied to one place continuously, as in the 
Lukens tests, but was distributed over the middle of the piece, causing 
a gradually increasing curvature until the desired radius was obtained. 
One-half of the test pieces were bent to a diameter equal to three times 
the thickness and one-half to that of twice the thickness. Two pieces 
of J-in. material broke because too sharp an angle was formed. These 
bends were very satisfactory, and proved that, by careful manipulation, 
almost any desired degree of bending could be obtained without 
fracture. 

These results are shown in the photograph, Fig. 1, Plate XV. 

It was not expected that the nickel steel would stand the same bend- 
ing as the carbon steel, but it was believed that it would stand a suffi- 
cient amount to prove its toughness or lack of brittleness. 

Wherever in the tables, in columns headed "Angle of bend" a num- 
ber in parentheses appears, the unenclosed number is the angle to which 
the material bent when the crack first developed. 

The angle of bend and the radius of the bend were obtained from 
sketches made by tracing carefully the outlines of the bent specimens. 
The radius was found by deducting, from the radius of the outside 
surface of the bend, the thickness of the specimen. The radius of the 
outside surface could be determined easily and with fair accuracy by 
being taken in a plane midway between the edges. Cracks usually 
developed first in this portion and spread to the edges. The testing 
was done in March and June, 1906. 

Bending Tests, Punched, Reamed, and Punched-Riveted Specimens. 

The bending tests on nickel and carbon steels, like the corresponding 
tensile tests, were arranged to show relatively the effect of punching 
material full size, of sub-punching and reaming to full size, and of 
riveting punched work. The full-sized holes were yS in- in diameter 
and the sub-punched holes, \^ in. in diameter, i in. of material, there- 
fore, being reamed away. 

The test was not included in the original scheme of testing, hence 
the specimens had to be cut from the excess material. None of the 
|-in. plate of nickel steel was available, and only small pieces of the i 
and |-in. plates. For nine of the test pieces of nickel steel, it was 
necessary to use the pieces 4 by J by 12 in., designed and used for test- 
ing rivets in single shear. The greatest load on any joint was 63 400 
lb., or 31 700 lb. per sq. in. This did not exceed the elastic limit, which 
from Table 30 was 62 000 lb. An examination of the results in Table 
37 does not furnish any evidence that the pi-evious use had worked an 
injury to the material. The location of the specimens in the plates 
seems to have had no influence on the results. 



NICKEL STEEL FOR BRIDGES 233 

The specimens were machined to a width of 3 in. with these excep- 
tions: Pieces BPRL 65, BPPL 78, CBPRL 65, CBPPL 78, and 
CBAPPL 75 were 2{;] in. wide. The punching, reaming, and riveting 
were conducted by the usual shop methods. The diameter of the die 
used in punching was i^- in. larger than that of the punch. The same 
riveter as for the tensile tests, and the same scheme for protecting the 
specimens by a rV'in. plate on each side, were used. The metal under the 
head of the rivet was bright, and, no doubt, was compressed as in the 
tensile-test specimens. As comparative results only were sought, there 
was no experimenting with mandrels of different diameters. The pieces 
were all bent, as described under the tests of plain specimens, in the 
testing room of the Lukens Iron and Steel Company, with a mandrel 
rounded on the end to a diameter of 2 in., a continuous pressure being 
slowly applied at the middle over the hole. 

The edges of pieces and of holes were smoothed by filing. This was 
overlooked for the first three pieces tested, namely, CBPRL 65, 93, and 
98 ; consequently, two pieces of nickel steel, BPRL 56 and 65, were 
tested in the same condition for comparison. The filing does not seem 
to have aided the carbon steel, but its effect on the nickel steel was 
marked. The two filed pieces bent 104° and 108°, respectively, and the 
unfiled pieces, 75° and 77°, with a corresponding difference in the 
radius of bend. 

The results are given in Tables 37 and 38, and in the photograph. 
Fig. 2, Plate XV. 

In every instance punching, sub-punching and reaming, and rivet- 
ing, alter the |-in. material more than the f-in. material of both nickel 
steel and carbon steel. The |-in. plain specimens bent without cracking 
through the greater angle; but, in these tests, the reverse was true. 
The difference between the effects of punching and of reaming upon 
the |-in. nickel steel is greater than it is upon the carbon steel, but, 
upon the |-in. material, it is less. 

Riveting increases the effect of the punching on both thicknesses of 
both steels and in about the same proportion. As a whole, these conclu- 
sions are the same as those drawn from the tensile tests. It could not 
be expected that nickel steel would bend through as great an angle as 
carbon steel. In the photograph, Fig. 2, Plate XV, the bent specimens 
are arranged side by side for comparison. 

These tests were made in March and June, 1906. 

Nicked Bending Test. 

The following pieces, NBL 27, 2i by 3 by 8 in., NBL 120, 2 by 5 by 
18 in., and NBL 189, 1 by 1 by 18 in., of both steels, were nicked i in. 
and fractured in two places by bending. The fractures were uniform in 
appearance and free from laminations, pipe, or other defects. The frac- 
tures of the nickel steel were finer-grained than those of the carbon steel. 



234 NICKEL STEEL FOE BRIDGES 

This determination of character of fracture was the sole object of 
the test. 

Drifting Tests. 

The consideration given this test, when the lay-out of tests was 
being prepared, was long and thorough. Specifications, for steel super- 
structures of all the large railroads, and of consulting engineers, were 
examined and tabulated. The specification of the drifting test seemed 
to be a purely arbitrary matter, as the amount of drifting a iM-in. 
or a TJf"in. hole, punched full size and spaced 1^ in. from a sheared 
or rolled edge, was required to stand without the edge of the plate or 
the periphery of the hole cracking, ran from 25 to 33%, 38%, 50%, 
and, in four instances, to 100 per cent. The other standard require- 
ments of carbon steel for bridges are given later. 

In the arrangement of holes finally decided upon, almost every 
recent specification, even the most severe, has been included. This ex- 
plains the irregular grouping of holes. The number was limited by the 
size of the coupon selected, namely, 12 by 12 in. 

With but two exceptions the hole required to be drifted was a, 
punched hole — these two specified a tI -in. hole reamed from a diam- 
eter of rh in. Holes of this sort were included, also holes drifted 
from the solid. The f-in., |-in., and 1-in. plates were selected, the first 
two because of the great bulk of structural material represented there- 
by, and the 1-in. plate, because it was proposed to use nickel-steel plates 
of this thickness in a large bridge, plans for which were being prepared. 

Only one specification was found in which a taper for the drift-pin 
was specified, yet this is as important as the size and location of the 
hole. A pin with a large taper bulges the metal around the hole, and 
the edge cracks because of this, and not because the limit of enlarge- 
ment has been reached. The ideal pin is one with a scarcely perceptible 
taper. A taper of 1 in 12 was adopted for this test, and several pins 
of this kind were made of specially hardened blue-chip and Sanderson 
steel. 

The preparation of the test specimens and the testing were done at 
the Pencoyd Iron Works, the drifting being done by driving the pins 
with a heavy steam hammer, first through one side and then through the 
other, about J in. at a time. In this way excessive bulging was pre- 
vented. It will be noted that the largest cracks and the greatest num- 
ber of them are always on the side where the bvilging is the greater. 
The holes were all carefully laid out, but, in punching, a slight shifting 
of some took place; the amount is not large enough to be of import- 
ance. Burrs were removed from the edges of the holes, but no filing 
was done. The diameter of the die used was tV in. larger than that 
of the punch. 

Tables 39, 40, and 41, together with the photograph, Plate 
XVI, show the effect of drifting on specimens drifted according 



PLATE XVI. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1103. 

WADDELL ON 

NICKEL STEEL FOR BRIDGES. 



DRIFTING TESTS 

ON SPECIMENS SHEARED FROM 12 UNIVERSAL PLATES 



DRIFTEU ACCORDING 
TO d^LUinCATIONS 



DRIFTED 
TO CRACKING 




^ o^o'^l < o^o^i 



o 



C f f) 



'.:!f 




■f O 







Q 

2 



o 



w 



1^ 






HORIZONTAL EDGES 
VERTICAL 



PIECES 
SHEARED 
ROLLED 



I 2 



I 2 



HOLES A E F G K L IN PIECES NOS. 20 AND 113 PUNCHED % 

B C D H I J 20 113 

M P Q R V 21 114 

U 21 114 

N S T 21 114 DRILLED 

A TO V 164 165 

DRIFT PINS TAPERED I IN 12 



REAMED TO 'V,, 
FROM SOLID 



NICKEL STEEL FOR BRIDGES 



235 



to the requirements of the specifications mentioned, and Tables 42, 43, 
and 44 and the same photograph show the effect of carrying the drifting 
until cracks were developed at the edge of plate or the periphery of 
the hole. The nickel steel drifted with greater difficulty than the carbon 
steel. Through an oversight. Hole K, in BS 164, was not drifted as 
much as intended. 

An examination of the tables shows that the carbon steel drifted 
better than the nickel steel. The difference is not marked in the tests 
made according to specifications; but in the tests to cracking, compar- 
ing the enlarged diameters of the holes, the f-in. nickel-steel plates 
drifted 58% as well, the f-in. material, 72% as well, and the 1-in. 
material, 67% as well as the carbon-steel plates, but the carbon steel is 
cracked to a larger extent than the nickel steel. It is probably fair to 
conclude that nickel steel will stand about 70% as much drifting as 
carbon steel, and satisfy the drifting requirements of many specifica- 
tions written for medium steel. A hole spaced, as in the usual prac- 
tice, from a sheared edge, will not stand drifting to the same extent 
as one spaced near a rolled edge. These tests are too few to give any 
conclusive figures, but probably the ratio is nearly as 1 to 2. Punched 
holes seem to drift as well as, if not better than, either reamed or 
drilled holes. The percentages giving extent of enlargement (Tables 
42 and 43) may be arranged to show this as follows: 

TABLE 19. — Percentage by Which Holes Were Enlarged. 



Hole edge. 



3/8- in 
3/4-in 
3/8-in 
3/4-in 



Material. 



Nickel steel. 



Carbon steel. 



Punched. 


Sheared. 


Rolled. 


57 


47 


.50 


50 


53 


80 


(iu 


87 


(3 


110 


73 


517 


80 


133 


: 67 


67 



Reamed. 



Sheared. Rolled. 



Drilled. 



Sheared. Rolled 



CO 
73 
100 
113 



These tests were made in April and May, 1906. 

Close-Punching Test. 

The close-punching test is rarely specified, because it furnishes but 
little information about the physical properties of material not obtained 
by other and more usual tests. It would show up brittleness or exces- 
sive hardness. In this series its value was in causing to be punched 
at one time, a sufficient number of holes in the nickel steel of different 
thicknesses to permit forming an idea of the effect on the machine and 
punches, and also to show the ability of the metal to stand close 
punching. 



336 NICKEL STEEL FOR BEIDGES 

A rather light punching machine in the shop at Pencoyd was used; 
and, while no part gave way, the |-in. material was too heavy for 
it. Twenty-eight rf-in. holes, and twenty-five it-iii^- holes were 
punched in each of the two pieces, CP 22 (12 by 12 by | in.), and 
CP 115 (12 by 12 by J in.), and about the same nuraber in the corre- 
sponding carbon-steel pieces. The arrangement of these holes and the 
results of the test are shown in detail on Fig. 1, Plate XVII. 

Two punches were broken, one by stripping and one by shearing, 
while punching the nickel steel. The steel in these punches was con- 
sidered exceptionally good, but the toolmaker was not familiar with it, 
and frequent breakages had happened in regular work, consequently, 
the blame cannot be laid altogether upon the nickel steel. At the time, 
the shop management regarded this work as more of a test for the 
punches than for the nickel steel. Before this, in connection with this 
investigation, considerable punching of nickel steel had been done at 
Pencoyd, with but one breakage, a lis-iii- punch. One punch and 
one die were broken in punching material for struts at Ambridge. 
Several holes, xf in* in diameter, were punched in |-in. nickel-steel 
plates, both in preparing test specimens and in preliminary experi- 
mental work. As regards the effect upon the punch, this is not a 
serious matter. The heavier punching machines now in the large shops 
would not be over-strained were nickel steel to be introduced, but the 
smaller ones unquestionably would be. 

The rapidity of punching would not differ from present practice 
were nickel steel introduced, though, at the outset it might, because the 
"punching" flies out with considerable force and with a loud sharp 
report, and men unaccustomed to this think there is danger. 

The punched hole in nickel steel is very different from that in 
carbon steel — its edges are clean and smooth, with no appreciable com- 
pression of the "punching." For this reason it is thought by some 
that punching is less injurious to nickel steel than to carbon steel. A 
discussion of this will be found under tensile tests and bending tests of 
punched specimens. 

The holes in the "close-punching" coupons at the left and top edge 
were spaced so that J in. of metal was left between the edges of adja- 
cent holes, r,i in. between the edge of the hole and the rolled edge, and 
I in. between the edge of the hole and the sheared edge. 

The diameter of the die used was jV in. larger than the punch. 

The partitions between the holes actually varied from t\ to 4 in., 
and between the hole and the sheared and rolled edges from A ^" 
•^ in. The actual distances are so much smaller than stipulated in 
any specification ever written that a variation of yV in. or more is 
unimportant. 

The photograph shows the die side of these plates. 

The effect of the punching is practically the same for both steels. 
Where the partitions between the holes in the nickel-steel plates are 



PLATE XVII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1103. 

WADDELL ON 

NICKEL STEEL FOR BRIDGES. 



n </» — Ki _ 

-i o o 

=" 3) f^ -I «^ 



■V»"b 


• 


j 






• • 


m 




• 

• • 


^ ••••%••» 



m" 



S </> 



-g 







NICKEL STEEL EOIl BRIDGES 



237 



shown broken down, this resulted from inaccurate punching; as ex- 
plained, the operator was nervous. The deformation of each hole by 
the succeeding one is alike in both steels, it occurs on the die side only, 
all holes on the punch side being truly circular in form; there is, how- 
ever, a slight depression of the walls between adjacent holes. There 
was no deformation of the metal between the other holes in these plates. 
This test was made in March, 1906, 

Hammering-Flat Tests. 

A clause covering the hammering-flat test can be found in one or 
two specifications. It is strictly a punishment test to detect brittleness. 
Material is rarely treated in this way in bridge-shop practice except in 
fillers; but, in riveted-steel pipe manufacture, the corners of many 
plates must be scarfed by cold hammering. 

The test was made on 2 by 8-in. pieces, marked HL 14 and HL 131, 
cut from the -|-in. and the f-in. plates, respectively, of each steel, under 
a light steam hammer. A small round was first used to localize the dis- 
tortion, but only for a few blows, most of the work being done by ham- 
mering the end held on the anvil. There was very little difference 
between the nickel steel and the carbon steel; the latter broadened out 
more, but the thickness at the edge was not less and the cracks were 
a little larger. Nickel steel is so stiff under treatment of this kind 
that the strain on the hammer is very great. The results are shown in 
Table 20. 

TAELE 20.— Hammering-Flat Tests. 





Original. 


Final. 




Material. 


Width, 
in inches. 


Thickness, 
in inches. 


Width, 
in inches. 


Thickness, 
in inches. 


Remarks. 


Nickel steel 

Carbon steel 


2 
2 
2 
2 


% 

H 


lit 

''IB 


33 
#3 


No cracks developed. 
Two small cracks started. 
Two small cracks started. 
One quite larfje crack. 



This test was made in the forge shop, at Pencoyd, in March, 1906. 
Shop-Tooling Tests. 

By shop -tooling tests is meant submitting the steel to all the shop 
operations necessary to its fabrication, and watching closely the results. 

Saiving. — It has been stated that the 8 by 8 by 1-in. angle was 
sawed across. A time test was not made. The saw was not injured. 
Probably it revolved at the usual rate, but was fed more slowly. 

Shearing. — All the usual sizes of plates may be safely sheared in 
any of the large shops — in these tests the 12 by 1-in. and a 36 by i-in. 
were the heaviest sections sheared. The strain on the shears is consid- 
erable, and probably, were nickel steel introduced, heavier machines 



238 



NICKEL STEEL POR BRIDGES 



would have to be built. The same time is required, whether nickel steel 
or carbon steel is sheared. 

Punching. — A discussion of punching was introduced under the 
close-punching test. No difficulties greater than getting the operators 
accustomed to the material are presented. There would be a slight 
increase at the outset in the breakages of punches, but, without doubt, 
the toolmakers would meet this extra demand upon them. 

Beaming. — Progressive shops, to-day, are adopting dry reaming 
with high-speed, steel tools, the only lubricant used being a graphite 
paste. Many shops, however, still retain the slow-speed, ordinary, steel 
drill with an excessive use of a lubricant, principally water. This sort 
of tool will not stand up with nickel steel : it over-heats, the edge 
crumbles, the tool binds, and the progress is painfully slow. With the 
special high-speed, steel tool revolving at great rapidity, nickel steel may 
be reamed at the same rate as carbon steel ; graphite is used more liber- 
ally, and the tool wears out more frequently. In the fabrication of the 
struts for compression tests, nickel steel and carbon steel were being 
run through a drill press, side by side. The behavior of the tools was 
carefully watched, though no time test was made. Some of the oper- 
ators, Hungarians and Poles, had trouble with the sticking of the tool, 
but not after they became accustomed to the material. 

Drilling. — A number of holes were drilled in specimens for drifting 
tests, riveted-joint tests, etc., and a special time test was also made. 
The general observations made under reaming might be repeated here. 
Ordinary steel tools will not last for any length of time, but the special 
steels now in common use may be used successfully. 

A piece cut from the 1-in. plates of nickel steel, marked SD 161, 
and one from the 1-in. plate of carbon steel, marked CSD 161, were 
used for the comparative test, with the results shown in Table 21. 



TABLE 21 
Medium Carbon Steel 



Drilling Tests, with Common ^ ^ -Inch Twist Drill. 

1 G 



No. of 
hole. 


Depth of 
hole, in 
inches. 


Time consumed. 


Speed, in 
revolutions 
per minute. 


Remarks. 


1 

2 


1 
1 


1 min. 17 sec. 
1 " 18 " 


U4 
144 


Drill was not altered. 



Structural Nickel Steel. 




The ordinary drill will stand up several hours with the usual carbon 
steel; with nickel steel, it gets excessively hot, turns blue, the point 
crumbles, and the clearance wears off. 



NICKEL STEEL FOR BRIDGES 



239 



In the drilling of Hole No, 2, after 56 sec, the drill was permitted 
to cool off before finishing the hole. Such a tool, however, would not be 
used in practice. 

TABLE 22. — Drilling Tests, with Cleveland High-Speed No. 10 

Steel (Blue-Chip) Tool. 
Medium Carbon Steel. 



No. of 
hole. 


Depth of 
hole, in 
inches. 


Time consumecl. 


Speed, in 
revolutions 
per niiuute. 


Remtirks. 


1 
2 

a 

4 


1 
1 
1 

1 


min. 51 sec. 
" 49 " 
" 49 " 
49 " 


210 
210 
210 
210 


Same drill used throughout— it 
was in perfect condition. 



Structural Nickel Steel. 




Same drill used in above test, 
after being ground, was used 
tliroughout — the point had be- 
gun to crumble at end of sec- 
ond hole and was in very poor 
condition at end of fourth hole. 



The speed was changed for the nickel steel, because of the danger of 
breaking the drill at the higher speed. The carbon-steel chips were 
steel gray, and the nickel steel, blue. No lubrication was used. 

The speed, therefore, with which the carbon steel was drilled was 
1 in. in 35 sec, and the nickel steel, 1 in. in 52 sec. A blue-chip tool, 
in ordinary work, would last half a day without sharpening, whereas, 
if used for 5 or 6 min. on nickel steel, it is necessary to sharpen it. 
This factor is the more important by far, as considerable time is wasted 
in the re-dressing. 

This test was made in the machine shop at Pencoyd, in September, 
1906. 

To obtain a comparison of all factors, a much longer test would be 
necessary; the foregoing, however, throws some light on the subject. 

Planing. — For planing, the edge planer at the Pencoyd plant was 
used, and a blue-chip steel cutter. An attempt was first made to deter- 
mine the depth of cut that could be made in the 1-in. plate material of 
nickel steel and of carbon steel. Coupons 4 in. wide and IS in. long 
were bolted rigidly to the planer bed. With the carbon steel it was 
possible to make a cut of i in., but the belt could not furnish sufficient 
power to cut this depth in the nickel-steel plate. A T,rin. cut was 
made, and, after a total depth of fV in. had been removed from the 
18-in. piece, the edge was burned off the tool. 

Successive attempts proved it possible to take a maximum depth of 
about ^\ in. of cut in the nickel steel; the surface, however, was 



240 



NICKEL STEEL FOE BRIDGES 



very rough and torn. Eight cuts, tV in. in depth, were next taken 
from each of the 18-in. plates, and the times noted, as follows: 







Time, in seconds. 








6H 
6 


6 


6 5M 6 5-M 
5M 6 6 (% 


5^ 
5^ 


5M 

5^2 


Average 5% 
Average S% 



CapboD steel 
Nickel steel 

The carbon steel apparently had no effect on the tool. The nickel 
steel, after eleven cuts, had burned the edge off. A cut of yV iu. 
would be probably the maximum depth that would be taken from a 
long plate, because of the destructive effect on the tool. 

It may be concluded, therefore, that nickel steel is more difficult to 
machine than carbon steel, and the amount of work done in a given 
time would be less, probably not more than half as much. The strips 
from the carbon steel were short in length, and steel gray in color; the 
strips from the nickel steel were longer, and deep blue in color. This 
is evidence of the greater power necessary to cut the nickel steel. This 
test was made in March, 1906. 

Pneumatic Chipping. — Pieces 4 by J in. were bolted securely by two 
bolts to a heavy I-beam. The workman was a regular chipper; he oper- 
ated a Boyer hammer, size ly^f by 3 in.; and the chisels were of 
standard size. The results are shown in Table 23. 

TABLE 23. — Eesults of Pneumatic Chipping. 



Nickel Steel. 


Carbon Steel. 


Cuts. 


Time required. 


Length of strips. 


Time required. 


Length of sti-ips. 


First cut 

Second cut 

Third cut 


3 min. 9 sec. 
5 " 6 " 

3 " 15 " 

4 '• 3 " 


16| in. 
13i " 
12i " 
llj " 


3 min. 53 sec. 
3 " 18 " 

1 " 35 " 

2 '• 50 " 


16i in. 
13J " 
18i " 


Fourth cut 


Hi " 


Totals 


15 min. 83 sec. 


54^ in. 


9 min. 5(3 sec. 


54 J in. 







The first two cuts in the nickel steel were made by a 
workman using a short chisel. 

The third and fourth cuts in the nickel steel were made 
l)y another man with a new chisel taken from the stock 
room. (Workmen do not like these chisels because of 
their length ; they think they cannot do as much work 
with them as with the short ones.) 

Material removed=:2.58 cu. in. 



All cuts made with new chisels. 
Chisel was not affected. 



Material removed=3.29 cu. In. 



Material. 


Time. 


Length of cut. 


Amount removed. 


Nickel steel 


15.55 min. 
9.9S " 


54.12 in. 

54.13 " 


2.58 cu. in. 


Carbon steel 


3.29 " '• 








On a Basis of 10 Min. 


Nickel steel 


10 min. 
10 " 


34.8 in. (64.V) 
54.6 " (100>^) 


1.6C cu. in. (.50%') 


Carbon steel 


8.82 " " (100%) 











NICKEL STEEL FOR BRIDGES 



2il 



Hand Chipping. — Pieces, 4 by ^ by 18 in., were bolted in the same 
manner as for the pneumatic-chipping test. The workman was an 
expert chipper of long experience. The results are shown in Table 24. 

TABLE 24. — Kesults of Hand Chippikg. 



Nickel Steel. 



Carbon Steel. 



Cuts. 



First cut. 



Time required, ^f^f^^^ 



5 min. sec. 



Cuts. 



5J4in. 



First and second cut. 



Time required. 



5 min. sec. 



Length 
of strips. 



Workmen rested 5 min. 



Workmen rested 5 min. 



Second cut. 



5 min. sec. 
3 " 22 " 
1 " 4 " 
1 " 33 " 



Totals 15 min. 58 sec. 



2^ 



I 5 min. sec. | 7 in. 

First cut was lOJ^ in. long, remainder was on 

second cut. 

Second and third cuts 4 min. sec. 5^ in. 

% in. was on second cut, and 4}^ in. on third cut. 



16>4in. 



Totals 14 min. sec 



201^ in. 



In chipping the second strip, the chisel 
was broken after 2 min. 32 sec; chip- 
ping }4 in. further turned edge of 
chisel. After 3 min. 22 sec. of chip- 
ping, new chisel was tried; in 1 min. 
4 sec, piece, was broken out of chisel, 
and, after another ] min. 33 sec, a 
piece was broken out of a second 
new chisel. 

Material removed = 0.97 cu. in. 



Chisel was not affected. 



Material removed = 1.24 cu. in. 



Material. 


Time. 


Length of cut. 


Amount removed. 


Nickel steel 


15.9" min. 
14.0 " 


16.25 in. 
20.12 " 


0.97 cu. in. 


Carbon steel 


1 24 " " 






On a Basis of 10 Min. 


Nickel steel 


10 rain. 
10 " 


10.2 in. (71%) 
14.4 •' (100%) 


0.61 cu. in. (69%) 
0.89 '• " (100%) 


Carbon steel 





All this work was done in the shop at Pencoyd, and, though the 
period of testing was short, a fair comparison may be drawn. It may 
be assumed that the time required to chip nickel steel would be one- 
half longer than for the usual carbon steel. 

These tests cover practically all the shop operations except milling, 
and, as this is similar to planing, the same conclusions may be drawn. 

It is not possible from these few tests to state even roughly what the 
additional cost of fabrication would be, should nickel steel be intro- 
duced. 



242 NICKEL STEEL FOR BRIDGES 

With the addition of one or two machines and a few men, the same 
number of pieces of nickel steel as of carbon steel could be handled in 
the shop, provided sections of the same size were used in the structures ; 
but if advantage were taken of the high tensile strength of nickel steel, 
to cut down the size of the sections, then the output in tons would be 
decreased very considerably without a proportionate decrease in operat- 
ing expenses. The general expenses of sales and auditing departments, 
of general office, drafting-room, template shop, and, in the bridge shop 
itself, of laying-out, assembling, riveting, painting, and dead labor 
would not be affected materially. 

Bearing-on-Pests Test. 

The bearing-on-pins test has considerable value for comparative 
purposes. It was designed to give the bearing values of structural 
nickel steel and carbon steel and of rivet nickel steel and carbon steel 
on a hard steel pin 1 in. in diameter. A special contrivance, illus- 
trated by Fig. 73, was built for this purpose. 

This contrivance was also designed to test rivets in double shear, 
which explains the presence of a l-in. hole over the 1-in. hole. The 
steel pin was turned to a driving fit, and was supported at four points 
to minimize bending. The two inner bearings are of tool steel, 
specially hardened and spaced with just sufficient clearance to permit 
easy withdrawal of the specimen after testing. 

The pieces tested were cut from the 12 by 1-in. Universal rolled 
plates of nickel steel, Heat No. 17 673, and of carbon steel. Heat No. 
33 342, three from each ; also three pieces from each of two similar 
plates rolled from a rivet grade of nickel steel and one of carbon steel. 
The following physical properties of the rivet material were reported 
by the mill : 

Kickel rivet steel. Carbon rivet steel. 

Composition Heat No. 2 096. Heat No. 19 !.'41. 

Nickel 3.30 

Carbon 0.17 0.13 

Manganese 0.53 0.43 

Sulphur 0.028 0.025 

Phosphorus 0.012 0.028 

Silicon 0.020 

Ultimate strength .... 84 200 lb. per sq. in. 54 100 lb. per sq. in. 

Elastic limit 51 800 " " " " 37 600 " " « " 

Elongation in 8 in. . . . 22.5% 30.5% 

Reduction of area 50.3% 59.9% 

The twelve specimens were each 3.83 in. wide, 4.50 in. long, and of 
the thickness of the material from which they were cut. Slight varia- 
tions in length were neglected, because they would not affect the read- 



NICKEL STEEL FOR BRIDGES 



243 



ings materially. The average distance from the top of the specimen to 
the top of the pin-hole was 3.83 in., but varied from 3.73 to 3.86 in. 

All testing was done in the testing room of Tinius Olsen and Com- 
pany. The load was applied by the movable head of a 100 000-lb. 
machine, and was read on the beam in the usual way; the compression 
was measured by a compressometer, reading directly to to hoxt in-> this 
delicate adjustment being effected by an electric contact ringing a tap 
bell.* The zero readings were taken under a load of 2 000 lb., in order 
that all clearances in the contrivance should be eliminated from the 
results. 

CONTRIVANCE FOR MAKING TESTS OF RIVETS IN DOUBLE 

SHEAR AND FOR DETERMINING BEARING VALUES OF NICKEL 

AND CARBON STEELS ON HARD PINS- 



-3^- 



.._3 







r-|--T 






V 1 






M 


i 


^ 


-^y 


r 






«^ 


-e— 





-Face 

1 Plate 6'x ji'-S'i' (tool steel) 
%"Hole(anIIecU 
l"nole(dnlled) 
/^^ 2 Plates 6"x s^'-s" 

(tool steel, plane to %) 
' 2 Angles, 4:1' :x. z"x %- 6" 






3K 



-^-% Hole 
(drilled) 



Nil 4 I 



-.-* 



Lli- 



1^4-- 



>\cVA- 



Face to bearing- 




All rivets and bolts, 'f diam., drill holes, use turned bolts. 

Open holes to be solid drilled after parts are finished, riveted and bolted. 

Tool-sleel will be furnished to shop. 

Other material to be taken from stock. 

Use no paint in assembling. 

Face backs of angles accurately at right angles to a thickness of J^" 

Temper tool-steel after holes are drilled. 

Fig. 73. 

Two projecting arms rested on the base of the apparatus and two 
against the under side of the specimens which projected below the pin 
for this purpose. Under each increment of load there was a yielding 
of the pin and supports, as well as a compression in the specimen above 
the pin, and, after the yield point had been reached, a distortion of 

* A description of this Instrument may be found in Olsen and Company's Catalogue, 
Part C, page 16. 



244 NICKEL STEEL FOR BRIDGES 

the pin-hole. This yielding varied with each test; the pin was always 
inserted in the same position. To determine accurately the amount of 
yielding of pin and supports, a piece of nickel steel with a half-hole 
only was used; one pair of arms, instead of resting against the speci- 
men, rested against the under side of the pin. A number of such read- 
ings were taken before, between the specimen tests, and after. The pin, 
after the series of tests on the plate material was made, was found to 
have been bent perceptibly, hence a second pin was used for the tests 
of rivet material. 

The half-hole piece of carbon steel was used only once, because it 
was so badly distorted ; and the readings, therefore, were discarded. 

Headings Nos. 1 and 5 were neglected, because so large a part of 
the deformation of pin and contrivance caused by this first loading be- 
came permanent that distorted readings would have resulted from their 
use. The readings are not accurate to the last figure. Headings Nos. 
3 and 4 (Pin 1) and Nos. 6 and 7 (Pin 2), taken consecutively, show 
that the probable error through inaccuracies of readings is very small. 

It will be noticed that after each loading there was an increase in 
the permanent deformation of the contrivance; therefore, in making 
the correction to the intermediate tests of specimens, this was taken into 
consideration. Heading No. 8 (Pin 2) shows that for some reason the 
conditions were changed, and that this is not the result of inaccurate 
micrometer observations is shown by the readings for Specimen 3N. 
The conditions became again more nearly normal for SO. 

It was impossible with the apparatus at hand to make corrections 
for every variable. The results of these tests are not scientifically 
exact, but for all practical purposes they are closely approximate. The 
corrected amounts of compression for each specimen were obtained by 
deducting, from the micrometer readings for the specimens, the mi- 
crometer readings on the half -hole pieces, Nos. 1 to 9, inclusive, in the 
following way: The plate carbon-steel readings are based on No, 2 
(Pin 1) ; the nickel steel on intermediate values between No. 2, and an 
average of Nos. 3 and 4. The rivet steels, liV and 10, on averages of 
Nos. 6 and 7; 2N and 2(7 on intermediate values between averages of 
Nos. 6, 7, and 8; 3iV" and 3(7 on intermediate values between Nos. 8 
and 9. 

Errors arising from this method affect only the amounts of com- 
pression and not the elastic limit, which was obtained directly from 
curves plotted from the original readings. The loading of the carbon 
steel was not carried as far as the plate nickel steel, because of the 
very large resulting compressions in the specimens, the curves becom- 
ing almost horizontal. The elastic limit is fairly well marked except 
for Curve BRP 116. 

Tables 45 to 49, inclusive, give in detail all the readings taken and 
also the corrected readings. 



NICKEL STEEL FOR BRIDGES 245 

Table 50 is arranged to show the essential features of each test for 
easy comparison. The figures in parentheses in the column headed 
"Elastic Limit" are those obtained from the corrected readings. As 
shown in the column of "Original Dimensions of Pin-Hole," the pin- 
holes in the plate specimens were 1 in. in diameter, while those in the 
rivet specimens were slightly enlarged. An examination of the read- 
ings of the plate steel revealed variations for loadings under 50 000 lb. 
not directly traceable to the contrivance — they were greater for the 
nickel steel. The pin-holes in these specimens were not left perfectly 
smooth by the drill. In the case of the carbon steel, this surface, after 
a small load, became as polished as glass, but the surface of the nickel 
steel still showed roughnesses, even after the maximum load had been 
applied. The entire surface, therefore, could not have had a bearing, 
and as the bearing surface varied, it could be supposed that the amount 
of compression over the pin, under equal loads, varied also, and that 
the extent of this variation would decrease as the load increased. The 
pin-holes of the rivet steels, therefore, were scraped and smoothed with 
emery. 

It is evident from Table 50 that the nickel steel has a very much 
higher bearing value than the carbon steel; both within the assumed 
elastic limit and beyond, the resistance to compression becoming greater 
relatively as the load increases. Under a load of 115 000 lb. per sq. in., 
the amount of compression is only one-tenth of what it is in the corre- 
sponding structural carbon steel. 

These tests were made in February and March, 1906. 

Bending on Pins. 

Specimens 1 in. square and 18 in. long were used. The surface of 
the plate on two sides was left and the two others were machined. On 
one machine surface the pieces rested on two supports 12 in. apart — 
the points of bearing being rounded hard steel. The load was applied 
at the middle through a steel casting, with the edge rounded to a radius 
of i in. and bolted to the movable head of a 100 000-lb. Olsen testing 
machine having an autographic attachment. The results were obtained 
from the pencil record. The movement of head within the elastic limit 
was at the rate of 1 in. in 12 min., and beyond this point 1 in. in 3 min. 
After the maximum load had been reached, the test was discontinued. 

The results are given in Table 25. 

The depth of the carbon-steel specimens being slightly Ics^^ than 
those of the nickel steel, to make the results strictly comparable, the 
elastic limit of the carbon steel should be increased by 25 lb. and the 
maximum load by 50 lb. 

The values for "elastic limit" in Table 25 are those obtained at the 
point at which the stress-strain ratio changes, as indicated by a devia- 
tion of the record from a straight line. If this point be determined 



246 



NICKEL STEEL FOR BRIDGES 



as was the yield point in the tensile tests, it would be 5 100 lb. for the 
nickel steel, and 2 570 lb. for the carbon steel, and for the elastic 
ratios, 54 and 50, and the deflections, 0.101 in. and 0.066 in., re- 
spectively. 

TABLE 25. — Bending-on-Pins Tests. — Structural Nickel Steel and 

Carbon Steel. 

Specimens, 1 by 1 by 18 in. — Supports, 12 in. between centers. 



Mark. 



Load per Square Inch. 



Elastic 

limit, 

in pounds. 



Maximum 

load, 
in pounds. 



Elastic 

ratio. 

Percentage. 



Deflection at 

elastic limit, 
in inches. 



Nickel-steel, 12-in. Universal Plate — Heat No. 17 673. 


BNPi56 


4 740 
4 530 
4 640 
4 630 


9 580 
9 450 
9 480 
9 500 


50 
48 
49 
49 




090 


BNPlhl 


0.080 


£iVP159 


0.075 


Average 


0.082 






Carbon-steel, 12-in. Universal Plate — Heat No. 33 342. 



CBNP 156. 
CBNP 157. 
CSiVP 159. 
Average. . . 



2 240 
2 460 
2 490 
2 400 



5 100 
5 2S0 
5 080 
5130 



0.044 
0.055 
O.OCO 
0.053 



The general conclusions would not be altered, however, the elastic 
limit for the nickel steel being nearly double that for the carbon steel, 
with a deflection only 1.55 times as great. As in the "bearing-on-pins" 
test, therefore, it would seem that the nickel steel was decidedly the 
stiffer material. The maximum load carried by the nickel steel was 
1.85 times that sustained by the carbon steel. 

Compression Tests of Struts. 

Twelve struts were designed for this test, three of nickel steel, 10 
ft. long, and three of same, 30 ft. long from center to center of end 
pin-holes, and the same number of each length of medium carbon steel. 

Each was built of four angles, 3 by 3 by f in., and two plates, 12 by 
§ in., laced on two sides with bars 2^ in. and | in., and had an area of 
cross-section of 17.44 sq. in. The ends were heavily reinforced to in- 
sure failure in the body of the strut. 

The design of these struts is shown on Plates XVIII ;ind XIX. The 
material was rolled by the Carnegie Steel Company, at its Homestead 
Works, and was fabricated in the shops of the American Bridge Com- 
pany, at Ambridge, Pa. The tests were made on the 2 160 000-lb. 
hydraulic testing machine of The Phoenix Iron Company, at Phoenix- 
vilie. Pa. 



PLATE XVIII. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1103. 

WADDELL ON 

NICKEL STEEL FOR BRIDGES. 



SS^'^StSm ' i lSSr^ flSSSlR 




iPI 



Fig. 1.— LoNG-CoLi'MN Tests. 




P'iG. 3.— Long-Column Tests 




Fig. 3.— LuXG-CoLiMN Tests. 



NICKEL STEEL FOR BRIDGES 



247 



It was intended that all material of each kind, except the rivet 
steel, should be from the same melt. This was not effected; but the 
record in Table 26 shows that the variations are not greater than would 
occur in any bridge member. 

TABLE 26. — Material Used in the Fabrication of Struts. 



Side plates, 12by% in 
Angles, 3 by 3by % in. 
Batten plates. 10 by % i 
Pin plates, 12 by 5^ in , . 
llkiby^in 
" 6by%in.., 
Lace bars, 2}^ by % in . . 
Rivets, % in 



Nickel steel, Heat No. 



17 673 

17 749 (1-10 ft. 17 065) 

2 116 (23) 
17 673 
17 749 
17 673 
17 749 (55) 

2 192 



Carbon steel, Heat No. 



41 520 



Stock. 



(1) 45 254 



(6) 60 161 



Composition. 



Heat No. 


Nickel. 


Carbon. 


Blanganese. 


Sulphur. 


Phosphorus. 


Silicon. 


17 673 


3.68 


0.41 


0.76 


0.020 


0.005 


0.046 


17 749 


3.52 


0.36 


0.76 


0.030 


0.012 


0.058 


2116 


3.50 


0.38 


0.70 


0.04 


0.03 


0.04 


17 065 


4.25 


0.463 


0.67 


0.014 


0.019 




2192 


3.28 


0.20 


0.62 


^.023 


0.010 




41 520 


none 


0.23 


0.56 


0.025 


0.025 


0.020 


45 254 


none 








0.028 




60161 


none 


0.19 


0.45 


O.OSO 


0.005 





Physical Properties. 



Heat No. 


Elastic limit. 


Ultimate strength. 


Elongation in 
8 in. 


Reduction of 
area. 


17 673 
17 749 

2 116 
17 065 

2 192 

41 520 
45 2.>4 
60 161 


67 500 

58 966 

68 500 
56 300 

44 20O 
37 100 
40 700 


112 600 

"92 866 
108 600 
77 600 

66 200 
61 300 
60 800 


16.5 

'iihh' 

18.75 
25.50 

29.25 
31.75 
30.25 


50.0 

■28.'4' 
41.3 
51.3 

56.2 
56.2 
60.8 



The bending tests on this material were very satisfactory. Those 
for the nickel steel are fully detailed elsewhere in this paper. Those 
for the carbon steel. Heat No. 41 520, bent 180° with an opening of 
i in. before cracks showed. Heat No. 45 254 bent 180° flat without 
cracking. 

The material was received at the shop during January and Febru- 
ary, 1905. Shopwork was proceeded with during the latter part of 
April. 



248 NICKEL STEEL FOR BRIDGES 

"De Pontibus" specifications for medium-carbon steel were to 
govern the fabrication, the detail drawing, however, was not so marked, 
so the punching was started full size. The 6-in. pin-plates for all the 
struts, and twelve holes in one 10-ft. nickel-steel angle and three holes 
in another, were punched 1;' i"- in diameter instead of tm in. The shop 
also mis-sheared the nickel-steel pin-plates. These errors, indirectly, 
caused a delay of one year; and, before the rejected material could be 
replaced, all work was ordered stopped. The shop was ready to resume 
in March, 1906, but, during the interval, the rivet rounds, one 10-ft. 
nickel-steel angle, and several lace-bars were lost. In order to pro- 
ceed, a 3 by 3 by i-in. angle from Heat No. 17 065 was used, and two 
carbon-steel lace-bars were substituted for each missing nickel-steel bar. 
These were placed at the ends of the struts so as to have a minimum 
effect on the results. The 6-in. plates of nickel steel were replaced, and 
the other mis-punched material was used. Additional nickel-steel rivet 
material was ordered. Carbon-steel rivets were taken from stock. These 
matters being finally adjusted, shopwork was again taken up in April, 
and pushed through without mishap. Extreme precautions were taken 
at every step to keep the two steels separate. 

One punch and one die were broken during the punching of the 
nickel steel. All reaming was done dry, with only graphite lubricant, 
and with high-speed blue-chip reamers. One reamer could ream to 
yf-in. about 60 in. of yf-in. holes in the nickel steel before needing 
regrinding; while, in the carbon steel, it could do four or five times 
as much. The riveting was done with a machine capable of exerting a 
pressure of about 35 tons per sq. in. 

This riveting work was all done at one time, on the same machines, 
the rivets being heated in the same fire. No variation was made from 
regular shop practice. All rivets were tested, but none cut out. The 
heads of many of the nickel-steel rivets cracked at the edge in making 
and in driving, but, on the whole, the appearance was very good. 

The struts were shipped to PhcBnixville on April 21st, where they 
were tested early in June. 

A special cast-steel bolster was made by the Phoenix Iron Company 
for holding the ends of the struts in position in the machine. The 
struts, therefore, were free to move vertically, and, in this direction, 
were the same as round ended, this being the usual condition in i)iii- 
connected bridge construction. The short struts were not supported, 
except at the ends ; but the long struts, in order to counteract the bend- 
ing from their own weight, were counterweighed at two intermediate 
points, 8 ft. apart. To the eye, these long columns, when in the ma- 
chine, appeared straight, and a stretched string showed a camber up- 
ward not exceeding in any instance more than 0.34 in., and two struts 
that showed such a camber failed by bending in the opposite direction. 
It would seem, therefore, that this initial flexure had no influence on 
the result. 



PLATE XIX. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1103. 

WADDELU ON 

NICKEL STEEL FOR BRIDGES. 




Fig. ].— Short-Column Test.s. 




Fig. 2.— Short-Column Test.s. 



NICKEL STEEL FOR BRIDGES 249 

As the load was applied to these long columns, flexure took place, 
the normal condition being resumed on the removal of the load until 
a permanent set was evident. 

The pressure was registered by a Shaw mercury gauge, graduated 
to 10 lb., equivalent to 32 000 lb. load on the strut. Intermediate 
values could be easily read. The accuracy of this gauge was examined 
and certified to by the maker. 

The amount of compression in the struts under each increment of 
load was measured by a contrivance, designed by Tinius Olsen and 
Company, consisting of a long arm (7.54 ft. for short struts and 27 ft. 
for long struts), one end resting immovably on one batten-plate, 
and the other carrying a pointer resting in a punch mark made in the 
batten-plate at the other end of the strut and turning on an axis, 
placed so that any variation in the length of the strut was multiplied 
five times and measured on an arc divided to hundredths of an inch. 
This becomes a very delicate apparatus, when the length between bear- 
ing points is considered. The movement of the head of the machine 
varied from yV to f in. per min. 

Strut No. 1, the first tested, was the short one of nickel steel, having 
the rff-in. angle substituted from a heat containing 4^% of nickel. 
The load was applied continuously, and the pointer of the gauge and 
the mercury column were watched carefully. This procedure was 
adopted for the first test, with some hopes that the elastic limit could 
be detected by the movement of the pointer or by the mercury column, 
or perhaps by both. There was no perceptible change in the rate of 
movement of either. 

The two mis-punched angles were also used in this strut, also car- 
bon-steel lace-bars to replace the lost ones of nickel steel, as described 
before. 

In succeeding tests, loads increasing in amounts were alternately 
applied and removed, and the movement of pointer noted. This method 
was very successful in establishing what loads produced a permanent 
set, also the amount of temporary shortening under these loads. This 
quantity is the sum of two changes, one resulting from a compression 
of the material and the other from the bending of the strut. It was 
thought that a succession of applications of loads in this way would 
weaken the strut below what it might be expected to carry, so that, in 
the beginning of the work, a greater difference between the successive 
loads was adopted. The number was gradually increased, as the maxi- 
mum strength of the strut did not seem to be affected thereby. 

In Table 27 the smallest appreciable permanent set was assumed as 
0.005 in. and as a "set" of 0.005 in. in 7.54 ft. corresponds to a set of 
0.018 in. in 27.0 ft., the load producing approximately this amount, 
0.015 in., is included for comparison. 

It will be noticed that in one instance only did a long strut of nickel 
steel have a set of this amount. 



250 



NICKEL STEEL FOR BRIDGES 



TABLE 27. — Summary of Results of Compression Tests of 

Struts. 





Nickel Steel. 


Carbon Steel. 


strut No. 


Load per square inch 
producing: 


Load per square inch 
producing: 


Ratio: 
Nickel steel 
Carbon steel 




Set 
= 0.005 in. 


Set 
= 0.015 in . 


Failure. 


Set 
= 0.005 in. 


Set 
= 0.015 in. 


Failure. 


Set 
= 0.(105 in. 


Failure. 



lO-FoGT Struts. 



1 


53 700 




68 500 

68 500 

69 200 
68 700 


39 600 
39 600 

37 300 

38 800 


38 900 






2 




.39 800 






3 


51800 
53 800 




38 900 






Average. 




39 200 


183 


175 











30-FooT Struts. 



1 


39 800 
42 500 


41600 


44 400 
47 200 
42 500 
44 700 






29 600 
33 400 

29 600 

30 500 






2 


16 600 

17 000 
16 800 


23 200 
20 400 
21300 






3 








Average. 


41300 




345 


147 









The reading of the pointer was discontinued after a marked perma- 
nent set was shown, or when the temporary shortening was so much 
as to exceed the limits of the contrivance; in three instances it was 
possible to watch the pointer up to the failure of the strut. 

The permanent set in two tests was not obtainable, one in the short 
strut of nickel steel, as already explained, and the other in the first 
long strut tested — Strut No. 1 of carbon steel. The movement of this 
strut disturbed the extensometer so that the readings taken are in doubt. 

The short struts failed by the buckling of the angles and side plates, 
in four cases at the middle panel, and in two cases at the panel next to 
the middle. The first carbon-steel strut crumpled upward, the others 
sideways. 

The long struts failed by bending. The distortion of the pin-hole 
was not appreciable. 

The amount of deflection at the middle of the long strut after 
failure varied from 0.64 to 2.60 in. 

These tests were conducted with great care, and the figures given are, 
it is believed, correct. They are based on data furnished by the Phoenix 
Iron Company as to the relation existing between the mercury column 
and the pressure on the piston of the machine. 

An examination of Table 27 shows that there is no evidence of an 
elastic limit. The total permanent set was exceedingly small, especially 
in the nickel-steel struts. The readings, however, are very uniform. 



NICKEL STEEL FOR BrxIDGES 251 

The compression in the 10-ft. sti^uts, under equal loads, seems to be 
about the same for both steels, up to the point where the carbon steel 
shows a permanent set; and it then becomes greater for the carbon 
steel. The amount of bending of the long struts is the same for both 
steels up to the point of fhilure of the carbon steel. 

The results are shown in detail in Tables 51 and 52 and in the 
photographs. Figs. 1 and 3, Plate XVIIT. The nickel steel in this test 
compares very favorably indeed with the carbon steel, the short struts 
being three-quarters, and the long struts one-half as strong again as 
those of carbon steel. 

These tests were made in June, 1906. 

Coefficient of Elasticity. 

The coefficient or modulus of elasticity is the ratio of the unit- 
stress to the unit-deformation, within the elastic limit of the ma- 
terial. As its value varies inversely with the deformation, it may be 
regarded as a measure of the stiffness of the material. The less the 
change in length under a given stress the greater is the coefficient of 
elasticity and the stiffer is the material.* 

Six specimens were prepared, three of nickel steel and three of 
carbon steel, as follows: 

No. 158 from the 12 by 1-in. plate of nickel steel. Heat No. 

17 673. 
B 1 from a 6 by 1-in. unannealed eye-bar of nickel steel, Heat 

No. 17 749. 
B 1 from a 6 by 1-in. annealed eye-bar of nickel steel, Heat No. 

17 749. 
No. C 196 from a 12 by 1-in. plate of carbon steel. Heat No. 

33 342. 
No. C 162 from a 12 by 1-in. plate of carbon steel. Heat No. 

33 342. 
No. C 109 from a 12 by |-in. plate of carbon steel. Heat No. 

33 342. 

The specimens were turned for a distance of 9J in. between fillets 
to a diameter as large as possible. This form was used because the 
construction of the Olsen extcnsometer made a round specimen de- 
sirable, and because the area of a turned roi;nd may be determined 
with greater accuracy than that of a rectangular section. 

The loading was applied very slowly, and all readings were taken 
with the beam balancing. Within the elastic limit, a given load may 
be sustained for an indefinite period, but, as the piece begins to yield, 
the stretching relieves a part of the load. Restoring it again increases 

♦Merriman. 



252 NICKEL STEEL FOR BRIDGES 

the stretch, and this condition continues until an equilibrium is es- 
tablished, when the beam will remain "balancing." 

The extensometer read directly to t'oVti in. and by means of a 
vernier to tu soo in. The first reading was always taken with a given 
load on the specimen — 4 000 lb. for the larger pieces, and 2 000 lb. for 
the carbon-steel pieces. In one of the tables filed for reference in the 
Library of the Society, however, all readings have been reduced to an 
assumed zero stretch with no load on the specimen. The values ob- 
tained are not scientifically exact. They are as nearly correct as the 
apparatus and careful observations could make them. In several in- 
stances two sets of values for the coefficient are recorded; the "a" 
results are more nearly in agreement and it is believed are the more 
accurate. With pieces, "B 1," the range over which the "V readings 
were taken is small, and, though a change seems to take place at 12 000 
lb., it is only apparent, and, had intermediate readings been taken, it 
is probable that a series would have been obtained similar to those be- 
tween 12 000 and 25 000 lb. In the case of the carbon steel, after the 
"V readings had been taken, the load was removed and a second 
series taken as a check on the first, because of suspected errors — 
through slipping of either the apparatus or the grips. The range for 
the carbon steel was necessarily small, and it is believed that a more 
delicate apparatus should be used for such material. 

The values for the two steels are: 

For the nickel steel.. £" = 30 075 000, 1-in. plate. No. 158. 

= 30 440 000, 6 " E. B.,B1, Original. 

= 30 420 000, 6 " E.B.^Bl, Annealed. 
For the carbon steel.. £^ = 28 940 000, 1 " plate, C 196. 

= 29 480 000, 1 " plate, C 162. 

= 29 840 000, f " plate, C 109. 

The coefficient for the nickel steel is slightly higher, but the differ- 
ence is not great; and, as probable errors in measurement of speci- 
mens or loads may be as high as 500 000, it may be assumed that below 
the elastic limit the two steels behave almost identically. Beyond the 
elastic limit, however, the yielding of the nickel steel is more gradual, 
and is smaller in amount. 

Had the entire range, up to the yield point, been taken in each 
case, the nickel steel would still be found to have the higher co- 
efficient, the difference being greater than before. 

It will be noticed that the elastic limit, as determined by this 
method, is considerably below the yield point. Curves plotted from 
these readings show clearly the relationship between the two. After 
the amount of stretch up to the yield point had been noted, the tests 
were continued and the specimens broken with the results shown in 
Table 28: 



NICKEL STEEL FOR BRIDGES 



253 



TABLE 28. — Results op Tests for Coefficient of Elastictty. 



Mark. 



Area, 

in 
square 
inches. 



Load per Square Inch. 



«! C R 
08 S O 



■= o o 

tHfttt 

a 



t2 CO 



Elastic 
Ratio. 



is" 



K^O-fe a 



• ft, 5 ft 



Fracture. 



Nickel Steel. 


158.... 0.760 
Bl.... O.Tii 
Bl.... 0.727 


36 800 
40 200 
35 800 


56 600 
62 300 
56 400 


104 600 
104 300 
102 800 


35.2 
38.5 
34.8 


54.1 
59.7 
54.9 


17.3 
18.8 
18.8 


47.9 
53.3 

47.0 


SCup. 

S 14 Cup Unannealed. 

S y> Cup Annealed. 


Carbon Steel. 


C196.. 0.754 15 900 
C1B3.. 0.590 11900 
C109.. 0.383 15 700 


23 500 
35 400 
28 700 


58 100 

59 000 
65 300 


37.4 
20.3 
24.0 


38.7 
43.1 
44.0 


30.8 
32.5 
28.0 


58.6 
59.7 
60.3 


S }4 Cup. 
S % Cup. 
SCup. 



Comparing the results in Table 28 with those obtained in the 
tensile tests of A. A. S. M. specimens, it is seen that there is a close 
agreement in the case of the nickel steel, Table 56; in the case of the 
carbon steel, the yield point, as determined above, is lower than that 
obtained from the A. A. S. M. specimens, and this could have been 
anticipated, as the speed of the machine was different. The method 
just described allowed the full deformation of the piece to take place 
before the given load was increased; while, in the previous tests, the 
increase in load was continuous. 

The tests of turned rounds also gave lower values for the yield point 
and ultimate strength, because of the removal of the rolled surface 
or skin. 

From these tests it is also seen how, under certain conditions, the 
ratio of the elastic limit or the yield point to the ultimate strength 
becomes, for the nickel steel, 50% higher than for the carbon steel. 

The very low elastic ratio reported in Table 31 for the 1-in. plates 
of carbon steel may be partly accounted for by the speed of the machine, 
having been 1 in. in 20 niin., for these pieces, and 1 in. in 6 min. for the 
other pieces. The tests made at the plant of the Luken's Iron and 
Steel Company, Table 33, show, however, that not all the drop may be 
accounted for in this way. 

It would be interesting to experiment further in this direction, and 
to obtain the amount of permanent set under given increments of load. 
Only two attempts were made to obtain data of this kind. 

The load was removed from piece No. 158 after 30 000 lb. (39 500 
lb. per sq. in.) had been reached and a set of 0.0001 in. was noted. The 
removal of load from B 1 (original bar) after 40 000 lb. (55 400 lb. per 
sq. in.) gave a set of 0.0018 in. 



254 



NICKEL STEEL FOR BRIDGES 



Further attempts were abandoned because of the danger of serious 
derangement of the extensometer and the interruption to the continuity 
of the readings. 

Tests of Eye-Bar Material. 

This material was rolled by the Carnegie Steel Company from the 
two 51-ton heats, Nos. 17 673 and 17 749, known as "ideal shape nickel 
steel." 

The following sizes were tested : 



No. 


Size. 


Heat 
No. 


Size of 
ingot. 


Weight of 
ingot. 


Size of Weight of 
slab. slab. 


1 
3 
4 


16 by 2 in. 
8 by 2 in. 
6 by 1 in. 


17 673 
17 749 
17 749 


25 by 30 in. 
22 by 25 in. 
18 by 20 in. 


12 080 lb. 
8 2U0 lb. 
6 060 lb. 


IS^by 14i4in. 
10 by 8 in. 
10 by 8 in. 


3 580 lb. 
1 900 lb. 
1 345 lb. 



The bars were forged and tested at the Ambridge Plant of the 
American Bridge Company. 

Forging of 6-in. Eye-Bars. — On June 14th, 1906, the 6-in. bars were 
taken into the shop. The eighteen months' exposure had rusted the 
surface considerably. 

The general method followed in forging these bars differed only from 
the usual practice in that the desired temperature was obtained more 
slowly. Along with the four nickel -steel bars, sixteen plain carbon-steel 
bars were heated and forged. 

Each bar was stamped with a distinguishing mark — B 1, B 2, B o and 
B 4 — for identification. From one end of each, two pieces, 18 in, long, 
were sheared for specimen tests, one piece being tested as rolled and 
the other after having been annealed with the forged bars. 

The furnace used for heating was fired with gas. The bars were run 
in at 11.08 a. m. in couples, B 2 on B 3 and B 4: on B 1, the carbon-steel 
bars being put in at the same time. 

At intervals, the temperatures of furnace and bars were obtained by 
the Le Chatelier electrical pyrometer, and, when this failed, by a 
Queen and Company optical pyrometer. 

First Head. — The temperature of the furnace at 10.30 a. m. was 
1 200° (650° cent.), at 11.20, twelve minutes after the bars had been 
entered, it was 1 115° (600° cent.). The temperature at the time of 
forging was not obtained, as, through a misunderstanding, the heater 
removed the bars before the stated time. After the rough heads (known 
as cobbles) had been formed on the pile of two bars, they were then 
rolled, still together, and then brought back to the furnace. Prac- 
tically all the upsetting was done at this operation, the second upsetting 
simply perfecting the shape of head. 

The second heating was carefully watched; the "cobbles" were 
entered about 12.15 P. M.; at 12.40 p. m. the temperature of the furnace 
was 1 960° (1 070° cent.) ; at 12.52 p. M., it was 1 975° (1 080° cent.) ; 




NICKEL STEEL FOE BRIDGES 255 

and after this reading the thermo-junction of the pyrometer separated. 
At 1.20 p. M. the temperature was 2 000° (1 090° cent.) ; at 1.30 p. m. 
2 100° (1 150° cent.) ; and at 1.35 p. M., 2 100° (1 150° cent.). Bars B 4 
and B 2 were then removed for final upsetting. The furnace was then 
forced to a higher temperature, reaching 2 200° (1 200° cent.) at 1.42 
p. M., when Bar B 1 was removed. At 1.45 p. m., Bar B 3 was removed, 
the temperature being about 2 250° (1230° cent). The head of Bar 
B 1 was dished in punching, which necessitated reheating and flatten- 
ing under a hydraulic ram. 

After the final upsetting, the bars were rolled singly and then 
punched, and the ears were sheared off, after which each bar was again 
rolled and laid on edge on skids until the entire heat was forged. 

The scale formed during the heating was removed by rapid hammer- 
ing both before and after upsetting. 

Annealing of 6-in. Eye-Bars. — In order, if possible, to obtain a 
variation in the heat treatment of the bars, they were run into the fur- 
nace in pairs, with the short lengths for specimen tests on top, as shown 
in Fig. 74. 

Bl B3 B-2 i 

B3 , '==' '=' > I Stack 

B2 J I End 

Fig. 74. 

Bars B 1 and B 2 rested on the rolls and were separated from Bars 
B 4 and 5 3 by J-in. square rods. The specimen pieces were placed 
about 4 ft. from the ends of the bars. 

The free circulation of the gases through the pin-holes and the pos- 
sible overheating of the head were prevented by placing caps, 4 in. 
square, over the holes. 

All temperature readings were taken with the electrical pyrometer; 
this was calibrated before being used. The gun of the pyrometer was 
inserted through small rectangular openings, about 4^ in. square, 6 ft. 
apart and 15 in. above the tops of the rolls and 13 in. above the eye- 
bars; the temperature of the furnace recorded, therefore, is that about 
1 ft. above the top bar. The place occupied by the bars was probably 
from 50 to 100° (10 to 40° cent.) cooler than that indicated. 

The bars were rolled into the furnace at 9.40 a. m., June 16th. The 
furnace had been operated continuously during the night, and was, 
therefore, thoroughly heated. 

The heating of Bars B 4 and B 1 occupied 3 hr. 23 min., and of 
Bars 5 3 and 5 2, 3 hr. 50 min. 

The probable average temperatures of the bars, when withdrawn, 

^^^^•- Bl 1500° (820° cent.) 

B 2 1450° (790° " ) 

B 3 1500° (820° " ) 

B 4 1550° (840° " ) 



256 NICKEL STEEL FOR BRIDGES 

Forging of 8 hy 2-in. Eye-Bars. — The same marks that were 
stamped on the 6-in. bars were put on the 8 by 2-in. bars for identifica- 
tion. These bars were heated in an oil-fired furnace somewhat in need 
of repairs, so that the heating could not be controlled accurately; the 
extreme ends were much the hotter. The bars were heated and forged 
by themselves. 

In the first head : 

Bar B 1 was in the furnace 74 min., 
" 7? 9 " " " " 76 " 

The average length of each interval required for upsetting, rolling, 
etc., was 6 min., or a total of 18 min. for each bar. 

Reheating was made necessary because of the great width of head 
relative to the width of bars, the excess being about 80 per cent. 

On account of the poor condition of the furnace, the bars were 
shoved in about 3 in. further than usual. 

In the second head: 

Bar B 1 was in the furnace 58 min., 
" B2 " " " " 59 " 

((no u ii a u nij a 







B3 B3 Bl 




Bl c 


CZD !=] [!=! 




1 




k'2 1 1 




B3 1 1 






8"x 2" Bars 



B4 



B4C 



Stack 
End 

16"x 2" Bars | 

Fig. 75. 

The average length of each interval required for upsetting, rolling, 
etc., of Bars B 1 and B 2 was 4 min., or a total of 12 min. for each 
bar. Bar B 3 was heated a fourth time, and the total time for upset- 
ting, etc., was about 21 min. 

The total time occupied with both heads was 3 hr. 27 min. 

Forging of 16 hy 2-in. Eye-Bar. — The 16 by 2-in. bar was heated 
and forged with five plain carbon-steel bars. The heating was done 
in the furnace used for the 8 by 2-in. bars, and the upsetting, etc., 
were done by the same men and machines. For identification it was 
stamped B 4. 

The first head was in the furnace 1 hr. 45 min., the second, 91 min. 
The forging of the first head occupied 22 min., in three intervals; the 
second occupied 30 min., in five intervals. 

The total time occupied was 4 hr. 10 min. 

Annealing of 8 hy 2-in. and 16 hy 2-in. 5ars.--These bars were 
annealed in the furnace already described. The heat treatment dif- 
fered slightly, a higher temperature being attained. 

At 10.25 p. M. the bars were rolled into the furnace in the order 
shown in Fig. 75. 



NICKEL STEEL FOR BRIDGES 257 

The length of time the bars were in the furnace was 5 hr. 25 min. 

At the time of withdrawal the temperatures of the heads of the 
bars nearest the door were, for Bar B 1, 1 500° (820° cent.) ; B 2, 
1440° (780° cent), and B 3, 1400° (760° cent.). The temperature 
of the 16-in. bar was not taken because of the breaking of the thermo- 
junction. Its temperature, however, was probably 1 600° (870° cent.), 
the stack end of the furnace being approximately 100° (40° cent.) 
higher than where the other bars were. The approximate average 
temperatures of the bars were probably as follows: B 1, 1550° (840° 
cent.); B 2, 1500° (810° cent.); 5 3, 1450° (790° cent.); 5 4, 
1 650° (900° cent.). 

Tensile Tests of Full-Sized Eye-Bars. — These bars were tested in 
the new hydraulic testing machine of the American Bridge Company, 
at Ambridge, Pa. The pressure was supplied by a short-stroke, two- 
cylinder. Deeming hydraulic pump geared to an electric motor. The 
capacity of the machine is 4 000 000 lb., the pressure being read on a 
Shaw mercury gauge. 

The only nickel-steel bars manufactured (within the writer's 
knowledge) up to the present time are those made by the American 
Bridge Company for the Blackwell's Island Bridge, New York City. 
With a knowledge of the difficulty in obtaining the yield point of 
nickel steel from the movement of the mercury column, it was agreed 
between the manufacturer and the engineer that an extensometer 
should be used for establishing this point, and that it should be 
assumed to be the load producing a permanent set of 0.025 in. in 20 
ft. This method has not given entire satisfaction, because of the 
unusually low elastic limits reported at times. 

It was intended, in this experimental work, to follow the same pro- 
cedure; but the results thus obtained when considered in their rela- 
tion to the readings as a whole and to the physical condition known 
as peeling are inconsistent. It is believed that in several instances the 
yield point had not been reached when the extensometer had recorded 
a set of 0.025 in. in 20 ft. 

At the yield point there is always a marked change in the amount 
of stretch when the load is increased uniformly, and, between the 
elastic limit and this point, the stretch becomes gradually greater for 
each increment of load. Just before and at the yield point, the mill 
or furnace scale begins to fall from the bar because of its unelasticity. 
This is a certain indication that the bar is stretching rapidly. 

In Tables 53 and 56, therefore, the results have been worked out 
on this basis, and not on an arbitrary assumption based on a certain 
permanent set. 

This method for determining the elastic limit or yield point was 
the same as that used in the tests of struts, the load in increasing 
amounts being alternately applied and removed, and the stretch and 



258 NICKEL STEEL FOR BRIDGES 

permanent set at each application noted by the extensometer. The 
readings are given in Tables 54 and 55, corrected, however, so that 
the zero readings for stretch and permanent set occur under no load. 

Actually, the pointer was set in advance of the zero mark, as it 
was found impossible to set the extensometer pointer exactly at this 
mark. It was also found, in several instances, that, after a small 
load had been put on the bar, the pointer returned so far as to pass 
the original setting, and in the case of the 6-in. bar, B 3, it passed the 
limit of the graduated portion so that all the readings had to be 
estimated by the eye. 

The several readings taken at two points within the elastic limit 
of the 8-in. and 16-in. bars, 5 1 to 5 4, Table 55, furnish some 
information as to the probable error of these results. It will be seen 
that differences of several thousandths occur. Such a series of read- 
ings for permanent set as occur for the 6-in. bar, B 1, where, between 
loads of 300 000 and 330 000 lb., there was no change, furnish another 
reason why too much importance should not be given to these results 
by themselves. 

The complete graduated arc measured an extension of but 0.5 in., 
hence the readings were limited to this amount, which is far too 
small, it is thought, for obtaining all the information necessary in 
determining accurately the yield point. The readings for each bar 
were discontinued only when the limit of the extensometer had been 
reached. 

Another objection to this procedure was the inability of the opera- 
tor to sustain a constant pressure sufficiently long for a yielding of 
the material to take place, as described under "Coefficient of Elas- 
ticity" determinations. This was especially the case with the small 
bars; it was impossible to reduce the speed of the machine to what it 
should have been for experimental work. The speed could be varied 
within certain limits by an electric controller in connection with the 
motor operating the pump; and it also varied with each increase of 
load. Just before the maximum load for the 6-in. bars was reached 
the movement of head was about 3 in. in 1 min. ; for the 8-in. bars, 
IJ in. in 1 min.; and for the 16-in. bars about i in. in 1 min. The 
6 and 8-in. bars broke with a uniform silky fracture, and an average 
amount of elongation and reduction of area; and the drop of ultimate 
strength and elastic limit below those of the specimen unannealed 
tests was fairly uniform, so that the annealing described was evidently 
properly done. The number of tests is too small to draw any conclu- 
sions as to the effects or forging and annealing tempornturos. the dif- 
ferences between the individual bars being very small. 

The fine crystalline fracture of the 16-in. bar wo\ild give reason 
for thinking that it should have been held for a longer time in the 
annealing furnace and at a lower temperature. The fracture was 



NICKEL STEEL FOR BRIDGES 259 

uniform, and gave no indication of maltreatment; but the elongations 
are greater than those of the smaller bars. 

Coefficient of Elasticity — Full-Sized Bars. — It was desired to 
obtain the coefficient of elasticity in tension from these tests. For 
this purpose ten readings with the extensometer were taken of the 
stretch in each of the 8 and 16-in. bars, five with a given load as small 
as practicable, and five with a load just below the elastic limit 
(Tables 54 and 55). The readings vary slightly, as already noted, but, 
after excluding the doubtful ones, an average was used for the coeffi- 
cient determination. 

The following coefficients of elasticity were obtained: 

8 by 2-in., 5 1 — ^ = 26 900 000 

« « " « B2 — E = 27 910 000 

u a u u B3 — E = 27 790 000 

16 " 2 " B4: — E = 2Q 770 000 

and, similarly, from the 6 by 1-in. bars, the approximate values: 

6 by 1-in., 5 1 and 5 4 — ^ = 27 500 000 
6 " " " B2 and Bd — E = 25 800 000 

These values are much below the 30 000 000 obtained from the 
1-in. round specimens cut from the 6-in. bar, B 1. 

The most reasonable explanation of this difference is that the areas 
used in the calculations are too large. These were taken with calipers 
after all loose scale had been knocked off, and were the same as those 
used in obtaining the physical properties. The actual area of cross- 
section is always less than the section measured, because of the heavy 
scale resulting from rolling and annealing. Assuming the coefficient 
of elasticity to be 30 000 000, as obtained from the specimens, the 
extensometer readings would indicate that the loads per square inch 
given in Tables 53 and 56 are about 9% too small for the 6-in. bars 
and 7% and 6% too small for the 8-in. and 16-in. bars. Some of the 
errors may arise from the extensometer readings, but the fact that 
there is a close agreement among the bars of the same size would 
point to the first explanation. This scale is difficult to remove. It 
may be as much as J in. thick on the heads; and a scale 0.04 in. thick 
would be sufficient to account for this difference. 

Tensile Tests of Specimens of Eye-Bar Material. — The tensile tests 
of the specimens cut from the 6-in. bars, B 2, B 3 and B 4, were 
made by The Riehle Brothers Testing Machine Company of Philadel- 
phia. The results are disappointing because of the failure of the 
Riehle autographic apparatus to work properly. The yield point had 
to be estimated from measurements obtained by a pair of dividers. 
The specimens from 8-in. and 16-in. bars, with the exception of 
EB 32, were tested at Drexel Institute on a 200 000-lb. Olsen machine 



260 



NICKEL STEEL FOR BRIDGES 



with an autographic apparatus. The two curves obtained for Speci- 
mens EB 11 and EB 21 are very different from the others at the yield 
point and, therefore, are marked abnormal. The variation in speed 
was not sufficient to affect the results materially. 

The yield point, ultimate strength, and elastic ratio. Table 57, are 
lower for the annealed specimens in every instance but one {EB 32 
broken by The Riehle Brothers Company). The elastic ratio of this 
piece is higher; but the yield point is in doubt, as no dividers were 
used. The elongation in 8 in. of the annealed specimens is less than 
in the unannealed in all but four instances. In two of these it is 
equal, and in the other two but 0.2% and 3.3% higher; and in the 
latter case, the fracture of the unannealed bar occurred near the 
end gauge mark, so that the stretch on one side was limited by the 
fillets. The reduction of area of annealed bars is lower in all but two 
instances — the two pieces cut from the 8-in. bar, B 3. 

The differences are very small, and the large elongations and 
reductions of both the annealed and unannealed bars are indicative 
of a good quality of material. 

The location of these specimen pieces in the eye-bar is a matter 
of interest. The center line of specimens cut from the 6-in. bars 
was located 1^ in. from the edge; from the 8-in. bars, two pieces were 
cut, one from the edge and one from the middle of the bar; and from 
the 16-in. bar only one 18-in. length was cut, and this was again 
sheared along the middle line, giving two pieces, 8 by 2 by 18 in., one of 
which was annealed with the bar. Pieces for bending tests were cut 
from the rolled edge, and the tensile pieces marked "2" next to these 
and 3 in. from the rolled edge. The piece marked "1" was cut from 
the sheared edge, so that EB 41 and AEB 41 were adjacent in the 
bars; the former being tested as rolled and the latter being annealed. 

The yield point and tensile strength of the pieces cut from the 
edge and middle of the same bar are different. Edge pieces, with the 
exception of pieces EB 22 and AEB 22, gave lower results. These 
differences are practically eliminated by annealing. 

TABLE 29. — Comparison op Results of Tests of Specimens. 





Specimens Cut from 8-in. and 16-in. Eye-Bars. 


Specimens 


Cut from 


Physical 
property. 


Middle. 


Edge. 


6-iN. Eye-Bars. 




Original. 


Annealed. 


Original. 


Annealed. 


Original. 


Annealed. 




60 100 
101 500 

58.70/0 
21.1% 
48.50/0 


53 600 

97 800 
54.90/& 
20.10/0 
47.50,6 


55 600 

98 500 
■^6.4% 
21.90/0 
49.7% 


52 900 

96 600 
54.8% 
21.00/0 
47.40/0 


60 600 
103 100 

58.70/0 
19.40/0 
51.10X, 


54 800 


Ultimate strength.. 

Elastic ratio 

Elongation in 8-in. . . 
Reduction of area.. 


100 000 

54.5% 
19.10/0 
48.00/0 



NICKEL STEEL FOR BRIDGES 261 

Bending Tests of Eye-Bar Material. — As already described, pieces 
18 in. long were cut from each bar before forging. One piece was 
annealed with the bar from which it was cut, but the other received 
no heat treatment whatever. 

The specimens for bending were taken from the edge, and one side 
only was machined. They were 2 by 18 in. by the thickness of the 
bar. 

The bending was done in the hydraulic bending machine at Pen- 
coyd. Pa., in the American Bridge Company's shop. 

A mandrel with end rounded to a radius of 2 in. was used for the 
specimens 2 in. thick, and the anvil was U-shaped with a mouth Si in. 
wide. The casting was broken in the first attempt. A similar, but 
heavier, anvil with a O^-in. mouth was used throughout the series. 

Each specimen was bent as much as possible by being forced into 
this opening by the 4-in. plunger, and then bent farther by pressure 
on the ends. It was desired that the radius of bend should be reduced 
to the thickness of the material; and in every instance the bending 
was greater than this. 

The specimens from the 6-in. bars were similarly bent, the plunger 
being 2 in. instead of 4 in. in diameter, and the aperture of anvil 
being reduced to 5| in., thus increasing the amount of bending about 
the end of plunger. No definite amount of bending was contemplated, 
therefore some of the pieces were bent to cracking. Table 58 and the 
photograph of the bent specimens shown by Fig. 2, Plate XVTI, give 
in detail the results of this bending. 

The radius of the inner surface of the bends varied from 0.4 to 0.7 
of the thickness of specimen, thus satisfying the usual requirements 
demanded of a medium carbon steel of 60 000-lb. ultimate strength. 



262 



NICKEL STEEL FOR BRIDGES 



APPENDIX A. 

TABLE 30. — Tensile Tests on A. A. S. M. 
Made on 200 000-lb. Olsen 



Test. 



Original. 



Section, in 
inches. 



Area, 



square 
inches. 



Load per 
Square Inch. 



■a ,-0 
r.9 3 

ftO. 








Percentage 
OF Elongation in: 


2 in. 


4 in. 


6 in. 


Sin. 


Per 

cent. 



O Pi 



Per 

cent. 



12-iN. Universal Plates. — 


Heat No. 17 673. 












Minimum 

Average of 4. 
Maximum . . . 


1.510 X 0.385 
1.510 X 0.385 
1.510 X 0.385 


0.581 
0.581 
0.581 




115 200 

116 300 

117 200 




25.0 
29.3 
32.5 


19.5 
20.8 
22.5 


16.7 
18.0 
19.3 


15.3 45.1 
15.8 46.4 
16.8 47.5 


Minimum 

Average of 4.. 


1.505 X 0.385 
1.505 X 0.385 


0.579 
0.579 


66 700 

67 875 


112 800 
114 550 


58.8 
59.2 


21.0 
26.0 


19.5 
20.7 


16.7 

17.2 


13.8 43.9 

14.8 45.2 


Maximum .. . . 


1.505 X 0.385 


0.579 


70100 


116 600 


60.1 


29.0 


21.5 


17.7 


15.6 


45.9-j 


Minimum 

Average of 4.. 
Maximum 


1.510 X 0.505 
1.510 X 0..505 
1.510 X 0.505 


0.7G3 
0.763 
0.763 


64 200 
64 200 
64 200 


110 600 

113 020 

114 700 


58.0 
58.0 
58.0 


30.0 
31.0 
33.0 


20.5 
22.7 
24.0 


17.7 
18.5 
20.3 


15.5 
16.4 
18.0 


45.2 
46.2 
46.9 


Minimum 

Average of 3.. 
Maximum 


1..500 X 0.505 
1..500 X 0.505 
1.500 X 0.505 


0.758 
0.758 
0.758 


61400 
62 060 
62 700 


106 700 

107 8.30 

108 400 


57.3 
57.5 

57.8 


26.0 
28.5 
30.0 


23.5 
23.7 
24.0 


19.0 
19 2 
19.5 


16.6 
16.7 
16.8 


49.1 
49.8 
50.7 


Minimum 

Average of 4.. 
Maximum 


1.510 X 0.750 
1.510 X 0.750 
1.510 X 0.750 


1.1.33 
1.133 
1.133 


61 400 
61830 

62 000 


106 000 
10() 900 

107 700 


57.2 
57.9 
58.5 


31.0 
33.5 
36.0 


23.5 
25.0 
27.0 


18.7 
20.3 
21.7 


16.3 
17.5 
18.3 


42.5 
46.6 
49.4 


1 test only 


l.i^OO X 0.750 


1.125 


62 700 


108 400 


57.8 


27.0 


24.3 


20.7 


18.1 


47.4 


Minimum 

Average of 4.. 
Maximum .... 


1.250 X 1.005 
1.250 X 1.005 
1.250 X 1.005 


1.256 
1.256 
1.256 


55 400 
58 550 
61 600 


106 800 

107 320 
107 900 


51.6 
54.5 

57.4 


37.0 
38.5 
40.0 


26.0 
27.6 
30.5 


21.3 

22.7 
24.7 


19.3 
20.0 
21.5 


45.8 
47.4 
48.2 


Minimum . . . 
Average of 2.. 
Maxmium.... 


1.250 X 1.000 
1.2.50 X 1.000 
1.250 X 1.000 


1.250 
1.250 
1.250 


58 400 
5.S 500 
58 600 


104 000 
104 (iOO 
104 600 


55.8 
.55.9 
56.0 


38.0 
39.0 
40.0 


28.0 
28.7 
29.5 


21.7 
22.0 
22.3 


19.8 
19.8 
19.8 


48.9 
49.1 
49.4 



6 BY 6 BY 1 


-IN. Angles. — Heat No. 17 749. 












Minimum 


1.510 X 0.750 


1.133 


60 700 


102 000 


59.5 


30.0 


L'4.0 


20.0 


18.0'40.8 


Average of 4.. 


1.510 X 0.750 


1.133 


62 250 


102 500 


60.6 


32.0 


24.6 


20.8 


18.6 42.0 


Maximum.... 


1.510 X 0.750 


1.183 


05 600 


103 700 


63.3 


34.0 


25.0 


22.0 


19.0 


42.7 



8 BY 8 BY 1-iN. Angles. — Heat No. 17 749. 



Minimum 

Average of 4.. 
Maximum 



1.250 X 1.019 
1.250 X 1.019 
1.250 X 1.019 



1.273 
1.273 
1.273 



68 300 

54 100 

55 000 


97100 

97 870 

98 800 


64.6 
55.3 
56.6 


28.0 
84.6 
41.0 


27.5 

28.7 
30.5 


22.7 
23.7 
25.3 


20.0'46.0 
21.3 48.5 
22.8 51.1 



Note. — This table is condensed from the original by giving, for each group of tests 
* Distance from nearest end gauge mark. f Fracture slightly crystaUine. 



NICKEL STEEL FOR BRIDGES 



263 



(PAIiT II.) 

Standard Specimens of Structural Steel. 
Machine at Drexel Institute. 




Character. 



Remarks. 



W Cup. 
H Cup. 
1 Cup. 



S. 1 Cup.t 
S. H Cup. 
S. Aner. 
S. H Cup. 

S. % Cup. 
S. 1 Cup. 
S. 1 Cup. 

S. 34 Cup. 
S. % Cup. 
S. 1 Cap.rr 



S. }4 Cup. 
S. li Cup. 
S-U Cup. 



S. i4 Cup. 



H Cup. 

n Cup. 



s 
s 

S. ICup" 



S. 14 Cup. 
S. % Cup. 
S. 1 Cup. 



2.2 1 in. in 6 min. to 
2.9 I ultimate strength, 

3.3 then 1 in. in 2 min. to end. 



2.5 



Edge. 



1 in. in 6 min. within yield point. Interior 
1 in. in 20 min. before and well I Interior 

beyond yield point, then 1 in. Edge. 

in 6 min. to end. Interior 



1 in. in 6 min. throughout. 1 in. in 
6 min. well beyond yield point, 
then 1 in. in 2 min. to end. 

1 in. in 6 min. within yield point. 

1 in. in 20 min. before and well 
beyond yield point; 1 in. in ti 
min. again; finally 1 in. in 2 
min. 

1 in. in (5 min, throughout. 

1 in. in 6 min. within yield point; 
1 in. in 20 min. before and well 
after yield point; then 1 in. in 
6 min. to end. 

1 in. in 6 rain., 1 in. in 20 min., 1 
in. in 6 min.; finally 1 in. in 2 
rain. See No. 52. 

1 in. in 6 min. within yield point. 
1 in. in 20 min. before and well 

beyond yield point; then 1 in. 

inG min.; finally 1 in. in 2 min. 

1 in. in min. within yield point. 
1 m. in 20 min. before and well 

beyond yield point; then 1 in. 

in G min. to end. 



Edge. 



Interior 



Edge. 



Edge. 



Edge. 



Interior. 



Yield point lost. 



Yield point well marked. 



Yield point fairly well marked. 
Yield point lost. 

Yield point fairly well marked. 



Yield point lost. 

Yield point well marked. 



Yield point well marked. 

Yield point well marked. 
" " poorly marked. 

Yield point well marked. 



S. = Cup. 
S. - Cup. 
S. = Cup. 



1.5 I 1 in. in 20 min. beyond 3rield point. 

2.3 ; 1 in. in 6 min. to end. 

3.5 1 in. in 6 min. within yield point; 
1 in. in 20 min. before and well 
beyond yield point; then 1 in. 
in G mill, to end. 
1 in. in G min. within yield point. 
1 in. in 20 min. before and well 
beyond yield point; then 1 in. 
in 6 min. to end. 




Yield point fairly well marked. 



S. ^3 Cup. I 3.6 
S. % Cup. ' 3.9 
S. 1 Cup. 4.3 



1 in in 6 min. within yield point. 
1 in. in 20 min. before and well 

beyond yield point; then 1 in. 

in 6 min ; finally 1 in. in 2 min. 

to end. 




Yield point fairly well marked. 



minimum, average and maximum records, as shown in the first column. 



264 



NICKEL STEEL FOR BRIDGES 



TABLE 31.— Tensile Tests on A. A. S. M. 

Made on 200 000-lb. Olsen 

All Specimens Lo- 





Original. 


Load per Sqcare Inch. 


Elastic Ratio. 


Test. 


Section, in 
inches. 


Area, 

in 
square 
inches. 


'Hi 


O C p 
« cS o 


£^4 
S teg 

Sag 

p-Sa 


?a| 


o _'a 

In 



12-iN. Universal Plates. — Heat No. 33 342, 



Minimum. .. 
Average of 4 

Maximum 



1.510X0.355 
1.510 X 0.355 

1.510 X 0.355 



0.536 
0.536 

0.536 



37 300 

38 800 

41400 



39 700 

40 800 

42 500 



63 700 

63 720 

64 400 



51.9 
60.5 



61.7 
64.1 

67.8 



16-iN. Universal Plates. — Heat ISTo. 41 520. 



Minimum 1.510x0.380 

Average of 2 1.510x0.380 

Maximum 1 .510 X 0.380 



0.574 
0.574 
0.574 



84 500 

35 350 

36 200 



35 900 

36 250 
36 600 



66 200 

66 750 

67 300 



52. 1 
52.9 
53.8 



54.2 
54.3 
54.4 



12-IN. Universal Plates. — Heat Ko. 33 342. 



Minimum . . . 
Average of 4 

Maximum ... 

Minimum . . . 

Average of 4 

Maximum ... 

Minimum . . . 
Average of 6. 
Maximum. . . 

Minimum . . . 
Average of 2. 
Maximum. . . 



1.510 X 0.475 
1.510 X 0.475 


0.717 
0.717 


36 500 

36 880 


1.510 X 0.475 


0.717 


37 000 


1.510 X 0.755 
1.510 X 0.755 


1.140 
1.140 


32 500 

33 420 


1.510 X 0.755 


1.140 


34 200 


1.250 X 1.000 
1.250 X l.Oi.'O 
1.250 X 1.000 


1 .250 
1.250 
1.250 


26 400 

27 380 

28 300 


1.250 X 0.995 
1.250 X 0.995 
1.250 X 0.995 


1.344 
1.244 
1.244 


35 700 
25 700 
35 700 



37100 
37 320 

37 700 



33 700 

34 050 

34 400 



27 000 
39 230 
31.'i00 



26 900 

26 950 

27 000 



63 000 
63 770 


57.6 
57.8 


64 300 


57.9 


61 900 
63150 


52.5 
53.8 


63 500 


54.9 


59 200 

60 850 
62 400 


43.3 
44.9 
46.4 


58 500 
58 600 
58 700 


43.8 
43.8 
43.9 



58.1 
58.5 

58.9 



54.4 
54.7 

55.2 



45.4 
47.9 
50.5 



45.5 
45.8 
46.2 



Note.— This table is condensed from the original by giving, for each group of tests, 
♦ Distance from nearest end gauge mark. 



NICKEL STEEL FOR BRIDGES 



265 



Standard Specimens. Structural Carbon Steel. 
Machine at Drexel Institute, 
cated at Edge. 





Pehcentage of 
Elongation in: 


3 eS "s 
01 ft 
OS 


Fracture. 




2 in. 


4 in. 6 in. 8 in. 


Character. 


d* 

.2«i 


Speed o£ brealting. 



48.0 

45.7 

47.5 


35.0 
35.7 

37.5 


30.3 
31.1 

33.7 


26.9 
27.9 

29.4 


55.0 
56.2 

.56.9 


S. Ang. 

S. Ang. 
j S. Ang. 1 
iS.^Cup. I 


4.2 

4.6 

4.9 


1 in. in 6 min to ultimate 
strength, then 1 in. in 2 min. to 
end. 



48.0 
49.0 
50.0 



35.0 


30.7 


27.0 


52.4 


:«.7 


31.0 


27.2 


55.3 


36. 5 


31.3 


37.5 


58.2 



S. Ang. 
S. Ang. 
S. Ang. 



3.2 
3.5 
3.9 



1 in. in 6 min. to ultimate 
strength, then 1 in. in 2 min. to 
end. 



50.0 
51.2 

53.0 

62.0 
53.7 

57.0 

47.0 
54.9 
60.0 



58.0 
60.0 
62.0 



35.0 
37.1 


29.0 
30.4 


26.3 
27.3 


54.2 
56.5 


38.5 


32.0 


28.8 


58.4 


41.0 
42.0 


34.3 
35.3 


31.3 
32.2 


57.1 
58.0 


43.0 


36.3 


83.0 


58.7 


41.0 
42.5 
44.0 


34.0 
36.0 
38.3 


31.0 
32.4 
34.0 


56.3 
57.7 
59.9 


42.5 
43.0 
43.5 


36.7 
37.0 
37.8 


32.6 
33.1 
33.8 


60.4 
60.9 
61.5 



S. Ang. 

S. Ang. 
( S. Cup. ( 
IS. i^Cup. (■ 

S. Cup. 

S. Cup. 
( S. Ang. / 
) S. Cup. f 

S. 14, Cup. 
S. H Cup. 
S. 1 Cup. 



S. Aug. 
S. 1^ Cup. 



1.3 1 in. in 6 min. throughout. 

2.7 1 •' -'6 " 

„ „ 1 " " 6 " 

•5.0 1 n. II « II u 



3.8 
4.3 

5.1 



4.2 1 in. in 6 min. within yield point; 
4.7 I 1 in. in 20 min. before and well 
5.0 beyond yield point; 1 in. in 

, min. again; finally 1 in. in 2 
min. 

4.3 1 in. in 6 min. within yield point; 

4.4 1 in. in 20 min. before and well 

4.5 after yield point; liually 1 in. 
! in 6 min. 



minimum, average, and maximum records as shown in the first column. 



266 



NICKEL STEEL FOR BRIDGES 



So 



B% 



8x8Xl-in 6> 


<6X Jin. 














1 


Angles.— A 


ngles.— 


12-in. Universal Plates. 


— 








Heat No. I 


;eat No. 




Heat No 


17 673. 










17 749. 


17 749. 












1^ 




S > g 


g !> g 


g t> g 


g >■ &S 


m > ^ 


i: ►t- 


S 


» -I — 


» <d — 


23 ■< — 


» <i =• 


P <1 3- 


5= :S 


B 

3 

3 

3 


en 




Qimum 
erage o 
ximum 


aimum. 
erage o 
ximum 


Qimum. 
erage o 
ximum 


limum. 
erage o 
ximum 


limum. 
erage o 
ximum 


2. ^^ 

3 P 

is 

3 o 








(-b • 


(-b . 


Ms • 


ns . 










; *. ; 


ifr. : 


*. • 


■ .*' : 


. _*. . 


• f^ 














'w 'oi 'ax 


',(>. *. '*. 


en en 


r 


B'CB 




o S o 


























o 


« 2: 


O 


XXX 


X X 


XXX 


XXX 


XXX 


X X 


X 


^J fc^ (-1 


o o 


^ ^ ^ 


o o o 


O o c 


o o 


o 


B-O 





o o o 








en en en 


CO CO 


CO 


CB " 


a 






§S8 












50 o to 


-I <I 


o o o 






*" 




p 


^ _L H* 


l-t t~L 


►^ h-^ K- 


^ ^ ^ 


OOO 


o o 


o 


Area, in 


ts ts is 

-3 K! ^ 




S ig g 


M to to 


-J -1 ^ 

en en o< 


en en 


§ 


sq. in. 






03 05 


iO CO *o 


cji o> at 


1-1 1-1 i-i 










CO N) I-' 


3; OS 05 
X Ol >t^ 


CTl C3 CJI 

05 1— 00 


05 OS o? 
OD C5 to 


03 03 o; 
en to o 


<J OS 
to CO 


25 


Drop of 
beam, in 


CO 


CO O JO 






00 i-* C: 


eo OD *-- 


2 S 


en 


8 S 8 




8 g 8 


8 g 8 


8^8 


o o 


o 


pounds. 


1-,*" 

z 2 
o w 

IS 






o o o 


!:i Hi o 


ti li o 


^ ^ 


- 


Ultimate 






Oi tt!^ OS 


en OS t4 


o ^5 CO 


OS *. 




strength. in 


8 8 8 


8 i i 


o o o 


o o o 


OOO 


^ s 


o 


pounds. 


g 2 § 


s g g 


8 g g 


OS en en 

MOO 


en en a* 
-I OJ en 


OS OS 

to o 


CJI 

00 


Elastic ratio, 
per cent. 


03 00 i(^ 


to ro -1 


o 05 ai 


-' -^ -" 


en en oc 


iC>. to 


'-' 


^ 5S Si 


i> <! 03 


IS 8 g 


S ^ =3 


o ^ en 


OS en 


en 


Elongation in 
8 in., per cent. 






o oo o 


en en en 


o en o 


-J CD 


o 








O O! o 










1 


fe ft SS 


S g Si 


CI i£^ i£^ 
O CO 03 


ij^ k4^ eo 

CO 1-1 -1 


en >^ iCi. 

O -J OS 


fe fe 


S; 


Reduction of 


o o o 


j» to to 


05 1-1 .^ 


05 O OS 


to CO en 


O -^ 


to 


area, per cent. 


TCOtBCO 

r-OOO 
c c c c 


5^X1 


CO CO po 


CO (B 


Uixjimisi 


CB K X K 


o 

B- 






poo 




3 3 3 C= 


3 3 B 3 


P 

3 




cr? 3-d 3 


9!) 75 crq crs 


T;"d JQW 






3-,- 


• « • -p 


' ; ■ 


— .— 




■-» 


a 


— ■- 






— .— 










9 


l(>. ts >-^ 


fe. N) t- 


rf^ iO 1-1 


it^ CO 1-1 


iCi. CO to 


*. to 


o 


Location. 


or o> 00 


Jl «D O 


rf^ 00 oi 


o to to 


en en 10 


CO CR 


o 


Inches.* 










ic^ccic^eo 


CO .£1- .U CO 


CO III. CO .«^ 


At drop of 




o coc 




3 « 3 K 


o_ o_ 

» 3 y, 3 


3;;^;7'3 


3^3 










X 




CB a (B ft 
P P P P 


lin^ 


3-(B =''D 
? O .=^ o 


S.5-gp 


.a pp. 5' 


5^3 


.^ 


1 in. in : 




b. ^ ^4.^ 






►-1*^*^^, 








oooo 


OOOO 


"cq-o 
3 CO 3 01 


OOO" 

01 «i » B 


ooo^ 
m CB CO 3 


°'='a 


c 


At break, 


CO n n m 


g 03 m cn 


to CD a 




(tana 














I in. in : 




P P P P 


■5 p p p 




p p p .B 


p p p B 


p p 3 


o 






« i 


j;« g 


"s 


H^ 


«1 




Kj 






P- 


2 as 


5£ C^ ^ 




(B 




(B 






«1 


-3 's: 


S3 3:o 


— 


^ 




p. 








=1? 


• p B P 

^5o 


■a 
o 

B' 


•a 
o 
5' 




2 
B 


a 




F- 




?j: 


c 


o 




o 


B 






p 


III 

ff S 5 


(B 


CO 




ft 


w 






£^ 


r^ 








X 






B 


o 

3 


















p 
















3^g 


















p 













CD P3 

o 

to .'''^ 

§ I 

o ^ 

o H 

9 ^ 

^ oo 

2 jij 



" s 



O C/3 

2 CO 

° 2 

3 ^ 

I H 



NICKEL STEEL FOR BRIDGES 



267 



p ^ 



! 2 > 



« 2 

5' » 



<W c; i W c 



Si >► 2 

5' ' ' " 
S 



S g 



en en C7I 



CDCDCD ta^(— *^— COCOtD 



XXX XXX XXX XXX 

ooo ooo ooo ooo 



to 


5C CO 


'-.> 


-1 


'<» 


'j^ 




'.»>. 


m 


v« 


w 






















»l 


-J -I 


o 


05 


05 


iO 


to 


10 


■**■ 


ri^ 


•^ 


^ 


K* h-* 


„ 


„ 


^ 


_ 


_ 


o 


o 


o 


o 


to 


S 2: 


,_l 


fc_i 


^_i 






^} 


en 


fn 


en 


4^ 


$ 














S 




^ 


^ -I 


ou 


OD 


o 


o 


■"" 


jj 


00 



ecosto cocceo J^eoeo ^^t^os 



O CC 00 02 CO o 



-1 CO OC {O -^ 



O-JO OiOO o» 

ooo ooo oo 






05 lO ■-' 
CO CO I-' 

S S 8 



® ^ ^ Ci 05 05 



o c;t rf;^ 



s 


en 


.c^ 




rn 


kC^ 




rn 


fn 


-.> 


OS 


en 


Ol 








X 


VXI 


-^1 











-.O *- X *. 03 



eo 
oo 


03 
CO 


" 


§§ 


^ 


a 


g 


g 


^ 


CO 


§ 


g 


3 


g 


8 


o 

o 


8 


8 


g 


2! 


g 


8 


s 


§ 


C5 


en 

00 


g 


g 


en 


en 
eo 


g 


g! 


s 


g 


§ 


s 


rf^ 


«. 


10 


« 


en 


a> 


to 


C5 


en 


en 


en 


ts 


- 


-■- 




' 


_._ 










-^ 





CKCOCC (/JCO- 



9^; 



>.t>o'^ ^oo> >o^'^ 



4^ CO to 

en oo bo 



en CO to en 4^ to 



CO o c ui it^ or 



CO 



co>-^co^^ t~i CO .^ WW 

g en„ en cn„ cn^ 

B* CO E' $ ft 5* c^ B" 

.D p .= rt p .B p .3 



_ enen- 
5 (? o .^ 



coMCek 



COh-.t-iH-. 1-i.i-ih-tCO i-J. I-* h-». CO I-* h 



U'ooi tnow- 



P n o o 



I.W li' t.W -, lU t.W kb M 

oocsp oooP 



x cn cn S 

IX » a r" 



en w B^co 

O f5 F O 



01 




5' 


o 

i 


^1? 

.W 5'S; 


1^ 


Ultimate 
strength, 

in 
pounds. 


a> 5 2" 


Percentage 

of elongation 

in 8 in. 


Reduc- 
tion of 
area. 
Per cent. 


O 

S3 

•i 


o 

C 


"* 


rfl 


D3»3 
>- H 

2 ° 
5 o 





o 



2 W 





P 


o 






!zi 








GO 


13 


^ 


tc 




> 


a 


h:1 






d 


t-< 


^ 


D 


M 




O 




O 
o 


o 


O 



p 




--* 




o 


cy2 






CO 

O 

GO 
H 
W 
cl 
o 

> 

Q 
> 
W 

o 

!z! 

IXi 
H 

H 
H 



268 



NICKEL STEEL FOR BRIDGES 



* iz; 
^ 2 
(0 ra 



>0 1^ 



^ 2: 



§. H. 



S ;3. 



E 1= 



Heat No. 16080. 




Heat No 


17 673 










sSOg 


2>g 


gt>g 


g>g! ^ 


t> 


s 


g> 


s 
















S 






M2 5. 
- p g 

S ■* a 
Bo .3 


S55. 

t^ <y i-t 

B0.3 


«2 2. 
p^3 


pi 1 

3 - = 





§■ 

c 
3 


«2 

3 












. r^. 






. Ma 








; jo; 


: .N>: 


t *^*. 


- I^- 


^ 




* f*^ 








««^ 


V^W^^ff^ 


^^SS ^iS^ « 


^ 


« 


^ 


^ 


Thickness, in 


inches. 


















it 






ISt9^s 


NifOiO 


fi tsto 


jototo to 


to 


to 


to to 


to 




oo 




ggg 












00 


p 


%i 




ags 


-Jostn 


-im*. 


00^-J =5 




























e 




ENSIGNS 
GINAL N 
ECTION. 


h-L|-Lt-L 


000 


t-i ^ ^ 


»-'-'" 








00 









-v>— 3 *5 


muioi 


en o-> en -J 


.} 


^ 


00 -J 


-} 






» 






















ifflCO-J 


«5 05W 


i0*.O 












■ CB B 






^^^ 


^^^ 


tStSiA 


^-*H-l "-l h-i 


,_^ 


^^ 


1-1 tJ 


^ 


5-^ 
& 




^^b 








50 o?0 CD 


:0 


-XI 


0:0 


°S 


Sf 




[MENSIONS 

ACT0RED St 

TION. 


(XOSrt^ 




Wl-'O 






























d' 




.•-r.^ 


000 






'-i 




00 

"5S 


p 


s £ 50 


'>&.'«. j^ 


« CO 05 "-J 


ggss 
















2^1 




O'^ i 


OC-.K. 




eooioi I*' 


to 










ososcri 


CSOOi 








™ 


Oi 





•a 


? 
















UiO 


C 


>■ 






m Qc 


oc;o 


>o 





Ul Ci» 


OI 


c g"2. 



























000 












f'fl 




















• CO 


















T3 ^<=! 


















timate 
rength, 

in 
ounds. 


g^g 




{D too 













<o 




^OJ4- 


05'-' so c: 











« 






1— ' M ^ 


-.} Oi J^ 05 


oc 


OS 


01 —J 


00 


a 








000 










000 


000 


000 


'-' 






^ 


m 


















"0 




















2- 


n 






(XlO<i 


0: Oi03 05 


0-. 


C5 


05 tJt 


ai 


^ 05 


5 


















S: 


i-ioocn 


-J MO 


UIOT 0-. 


05 iiH-i a> 






CO 


>«^ 


» 


^' 




















BT 


















,»■ 


r s 


ll^i^CO 


-J~i-} 


cuts to 


OI*.*. *^ 


CO 


to 


C5Cn 


UI 


i-i. 


2S 


mooi 


tntso 


0100 


uios m 


>«<■ 


en 


01 S5 







S 


























































00 




catoi-^ 


|6.>^C0 


ta^ l_L l-i 


oso: to to 


to 


I-' 


COCO 


CO 




COUIQC 


Ji-J-OO 


ceo w 


&50CX) Ol 





o< 


en CO 







.3§ 


















>V 




^^ 


^^^ 




^^^ 






tOi-' 


,_^ 


•0^0 


^ 


oso-j 


05 0: » 


01 Oi*- 


05 to^^ 








05 


ft 


io.coic 


-laii*^ 


OI0 05 


-J XOD .*- 


05 




CO 


>C> 


l-^^l 


1= 



00 


WUiir^ 


OQO 


OOC-'- 


_,, 





33 72- 


_. 






CO M 

5'S' 


05 Jq 


•1 •^ T 

CB K OJ 

£E£ 

5' 5' 5' 


rystallin 
L. Cryst 

R. J^2 

Crystal! 
L. Silky 
rystallin 
rystallin 
rystallin 









."' .f" 




ft ft (t 


WS 


iVwp 




0^ 













;2:S 




<^ 




?"' 





















l-b 





























>► 


H 




!2j 


K 







H 




C/5 




W 


t/< 








t-l 





m 










m 

y^ 










^ 





V. 


> 


!z! 




y 




^ 




C/J 









w 




1.J ' 


> 


g 


13 


g 


t?3 


cc 


M 


!?; 


p 





(/J 


rt- 








H 


Ca3 














M 


X 


a:|o. 


!2; 


<x> 








M 


.^ 


1— 1 


I2! 


-^ 


s 

X 


w 




(-f. 


H-t 


•* 


C 


> 






H 





NICKEL STEEL FOR BRIDGES 



269 



•S. 3d 



33 £ 



si 3 
5 * :? 



000 



boco'ui 



a > g 

5' ^ 3 



2 > g 
5' » 3 

3 o 3 



=■33 

5 J 5 



i^ i^ 1^ o5s 



tc to to 

'o b p 

-1 ot iK 



JO to to 

b "0 p 
o> en C 



^-« 




^_^ 


^ 




















05000 








-> 


-J 






_^ 


M 














a;. 






w 










to 




to 


LU 


4^ 




Of) 


'-7 


05 


"00 


00 


'nr 


TT 


be 


3D 


«0~J32 


to 












CC 






,_L WJ 1-1 


l_^ 


,_^ 


_^ 




















it^-P-CO 


,_i 








CT 


m 


01 


s 


vj 


m 




















**"'^ 


to 


~} 


>(^ 




UJ 




-J 




■^ 



1-1 O O i-> 5D 00 

O C5 o o ot o 

o to o 000 

000 o c o 



ci CJT in c; o C5 



O »! O 





UI 


en 


en 


C2 







O! 


OS 


en 




-< 


Oi 


en 















SDOCC 


CO 


^^ 


*~^ 


05 


-J 


to 


-J 


w 


ai 






















o:e;iM 


>^ 


OS 


^^ 


a 


X 


OD 


to 


— ^ 


^^ 


OOCi 








-I 


cn 


"■^ 


en 





en 


^ 



o: en en 
s -1 en 









to 


to 


to 




to 


to 


to 


00c OS 


»i^ 




05 










OS 


en 


ib.>(>.;0 


""" 





to 


03 


CO 


to 


OS 


-I 


OS 



O CCCCCOhBOT OOCOCCjOCBCCZiM 



B p B .:, 
J<3 or? cr? Q 3 















Thickness, in inches. 




H 




Dimensions of 

Original Net 

Section. 


m 


si 

5' 


1^ 



Dimensions of 
Fractored Sec- 
tion. 


B-g ?> 
235- 


Yield 

point, 

in 

pounds. 




>■ 

a 

►-I'd 

a 

Wcc 

<o 

d 


4 
Ultimate 
strength. 

in 
pounds. 


3:^ 

Ct c^ 

5-' n' 


4 in. 


Percentage op 
Elongation in: 


00 

a 


Reduc- 
tion of 
area. 
Total. 
Per cent. 






i-tB- 

P J? 
P 

11 

■ a 





td 



w 

m 
CO H 

d SB 

? I 

a >► 

t25 si 
o 2 

pi 



OHM ^ 

O -.!^- t^ 

O K 

O > 2 






?= 



C Hm t^ 

5 ai 



o 

00 

O 
» 

O 



'^v! 



270 



NICKEL STEEL FOR BRIDGES 



(D O W 

*§ I 



^ s: 



f* (N 



<« 3- 



g!>g 


£>s 1 


p < =■ ^ < — 1 


gCf<5 c 
C (0 s 


-i ^ i. 




3o- 


3o» 






■ .''^; 


to; 


ig;^« ^^i; 


to JO to 


JO to JO 


ooo 


ooo 






^^w 


ooo 


m cji oi 








0=->Q0 




JO to JO 


tOH^W 


ooo 


5D0 






^^^ 


OOO 


Zi'Vltb. 


o-a-ct 






OOMOJ 








tooci*^ 


»oai 




O CJl o 


ooo 


°^^ 




^^^ 




ooo 






CO JO CO 


oin tc 




OOO 










-J-l-J 




Wi-« O 


-<t-^Oi 


XOO 


COIO>-' 


ffiOI*^ 


ocin 


ooo 


►-^o 


C«tO(0 


WO-5 


OOIO 






o: JO to 


^■occ 


-*50Cl5 


to-] 09 


•^ 


CO>T) 


^^1 




o 






' '<< 


•< 




» 



^ 



g;>g 


s>s 


so -4 S- 

M 2 :::. 


p ri a 


5' 5? 5 


5'^ 3 


C re S 


£?e 


So- 


3oS 






: *-; 


; jo; 


ig^^ 


^XS3Fv 


to to JO 


to to to 


ooo 


OOO 






l-Lt-J N^ 


OOO 


Cncr» oi 


<!-:!-l 






COOJGC 


oc oot 


to to to 


^1-1 1-" 




CO-^5D 


-^iOSOl 




h-i h-tH^ 


ooo 




05 cj» m 








o.*-o. 


>Cv«.^ 


>c>4^>«- 


4i iO.-> 




O C5 o 


cntoo 






oo5 






or or en 






1— L 1— 1 k-A 


»^ ic JO 


otoo 




ooo 


ooo 


CET.rC 


05 05 


00 wo 


tC -^ID 


4i — to 


50 0*. 






to — ^-^ 


— OtD 


OOOOI 


otoc« 


— 


o: oicn 


cji too 


COQCW 






l-'OO 


Oo-ioa 


OSOCII 


to — o 


o 


~^C0 


,,g 


r"r> 


X 

t 


7-7-w 


c 


O^ 


r ri 


I?' 


ui 


» 





Thickness, in inches. 



o C fO 
.-' CD g" 



=-^ 






7i ••' V 



^^^F^i 



H q m 



z s 5^ 

• w ^ 

o o 



B = p £. 



CO fH 

•a S" — 



o 5 



o e 



^ o I? o 5 
(t> E.85 a ^ 











K 


to 


tH 


o 




o 


GO 


o 


i_j ^ 


o 


W W 




O t?j 


o^ 




o 


^0 


n 


^ o 




o >► 




in S) 


^ 


g; 


W^ 


3 


':i oo 


CD 
r-l- 






M IH 












fD 


O 


X 

m 


t^ 








'X 


(— 1 


►fl 


^•^ 


H 




O 








g 




w 


ft> 


^ 



NICKEL STEEL FOR BRIDGES 



271 



-*■ * i 

o-u ■: 
a a 

l!l 

cr. 03 

» c '•" 
Cx o 
CO "^ 

OS'S- 

te 3 (T> 
go,. 

^ 6 

(I T 

O f 
OSS' 

tog 



•a 3" 



i-bie 
S 5 

"5' 
^5 



i3S*.co4^»ocioDUi-j~)coogc<o~J 






^ -} *» O J^ CI Ut CO *^ *»■ cc *-* ^ ^-* to _»o M 
l<i ^ bo »«^ o: K3 CO CI CO "o 5C "4 bx bs bo o cc 



CO 




Cc-CO O CO 






O O C = 3 3 


' 'S.'s ' 




-' -^ "k V! -" ai 

2 £.z;.s."c;i.E 






ai Tm X \ 1 » 1 
1 I 1 Pel P 


hole out 
across a 
hole out 
across a 


P 3-0 
O s 

p S.3 


cross — e 
across — 
ole outvva 
cross — e 
one aero 
across ai 
hole out! 


ward 
nd th 
ward 
nd th 


'3 5- ^ 


'2 S./.J5'2.2j5' 

Ec ||f.S 


,3" 3' 


.3 

'C 

TO 


(J? P-nr-^ 



to 

K 

W 



o 



5^ 

las* 

CO £,01 
tfl ^ to 

5' 2. 



> 



h^ 


fc3 


r; 


M 


!2l 


fei 


n 


o 






w 


^ 


O 




T 


H 




K 



M S 



H 


> 


O 


W 




!h1 




GC 


GO 


M 


c; 


W 


W 


63 


o 


a; 


^I) 


g 


Hh 


d 


w 


(S|Cn 


!z! 


!z! 




o 


«3 


!^ 




o 


b 





•^ 


h-( 


M|M 


1-/3 


>> 


e-JM 


H 






w 


H 


i25 


n 


H 




H 


W 


;^ 


d 


"^ 


!2! 


rt 


T) 


G 


> 


rl 




f 


!^ 


w 




n 


M 


y 


W 


>- 









o 


W 


e 




<1 


H 




5 


O 


H 


H 


H^^ 


M 


u 


=(■.1 


K 


b 




fi 



b M 



w 

o 



272 



NICKEL STEEL FOR BRIDGES 



:> 



* V. 


i^?iS!^i^(ij\tS\itSiSf!^i^«jr(«\4^3Sj|C^ 


> "• 








S ' 








S-" 


















2- ? 


to 






p. 


Q. (C 


^ , r #- 


,& o- . . . 


e » 


'•■•''" 


*■ HJ 




* f=l 


















"■„ 




















§g 


















1 










fK 




^- 






p 




















p< 








s- 


»-* — * 


1-1 t-1 1-1 ,-1 h-l .-1 1-1 V-1 U- 1 l-l 


1-1 ■ ■ 










»o 


s^ 










5D 






o 


* 












CD 








ffi 


to to 


1-1 1-1 H^ 




M 














* 


ODOOW 










s 


05 05 


Oi Oi Oi 




^ 








p" 








CD 
D 
C5 


CO 05 1-1 1-1 


1-1 to to to — -^ 1-1 1-1 1-1 1-1 1-1 o o o j 


VocnVDtDaoocno'MbDbi'-fc b>— tfiVotD^o j 










i-it-i 


l_l 1-1 1-1 h^ 1-1 1-L M l-L 1-L 




a 


' * ^- 1-1 




H-l 1-1 
























. . OS 






n 








? 
















S" 








P- 






o 


■d 






•3> 

o 


CD 






1 


a 






'O 


CD 




















p- 




, , 












p. 

CD 










p-js 








o ra 








sr "^ •■ ' 




-P 




<t> O - ' 


r-.-rrr*-*-'-*- »i» - ' 


■ O 








(t> 




r-P 




1 




^ B 




o 




Sp. 




c 












CLff ' •■ 




-^ 




• 1 
















B 




o. 




(W 








B- 




2 






D.C. 








oijonj 


'5 






















B Bs 


-B 






S.« 








3^?- 


B) 














(D 








& 





M 



o 



05 

OS 



o't 


1^ II 




glas- 


s's. B 


CD 05 CD 






O 


»• 


P 










CD 


1 


O 


l-B 


a- 


o 




a 


s-?^ 




(qB,| 




pB-a 




to 


B 
O 


X 


S" 


60W 




H jfc 




20 




-Sa 






•• to 






"- 


o 


" 








3 o P O 


s-as;:^ 


&^ 5-. .5. 




m S- TO 


(t 


W 


(D 


a 




»: 


T 


K 




w 







H 






t> 






W 






1— ( 










Ml 1-1 


W 


n 






W 




1 


ft! 
O 


> 




W 





S! 






M 


■*^ 


;/: 




K 


c 


^ 


M 


a 


o 




hd 


H 




n 


M 




o 


(A 

H 
so 




w 


O 







!2! 




H- 


rn 




o:(i- 


'Ti 




!_, 


w 




izl 






t> 


M 




!zl 


!2! 







M) 




td 


o 
»I1 




M 






>► 

^ 











M 

r! 




H 
O 


g 




Hi-i 


O 




doi 






^ 


« 
o 




M 






1— 1 


(/J 




> 


H 




^ 


W 









^ ?i 



H 


^ 


W 


M 


W 





o 


■^ 




^ 


lA 


H 


H- 


w 


ojcn 


a 


!z; 


o 




M 


O 


(* 


> 






!?! 


M 


o 


H 
W 





NICKEL STEEL FOR BRIDGES 



273 



tri >;^^a;^'5]tq:30to^ 



a" I 
ft "H 

,'' o I 

o ^ '■ 
c o 

3 „ 
Clo 
O (I 

a 2. 

o O I 

a 1 



?^ 

C o 

oS 

en a 
*S 
►< 3- 

la, 
^•? 



cc CO en 01 CT M CO CO e;i c;t e;i CO 



~ CO OO ^^ 1^ CO '-• 

cococo to o> tso: 



00 t3<?i:i^i^^ceM"^ii 

coococd otO{^a:o^ 



00 aoQ<cncj<cncocococo»:>{s 
Si Sioooooococoo u<cJi 



<«^cncoc 



j-i u. -} ai i-i CT ^ jj.^ en 

to M 00 00 to 00 X-\ 00 00 



W S5SJ cow CO 
c: 2 ^ (t — Q 



CooB 
ii a P 






•iz;cp 
o p 

n 2 



bS. 


3- 


Jf? . 









2>o 


(B 


rfif 


_ 


3- 


J^ 


- &» 




ce 


K! M 


O' 


(S S 




n> 


o 


"•o 




•O B<< 






pv- 





=. '3- 



o 




b>^«>,K^t:t;C!'s: 


tqtootDtt^ 










o 

!2! 


« 

8 








CO CO CJT en en CO 00 CO en CJ» OI M 1 


ts 




•\"^"^ 






0050505 0: C5a>o:c:osoJC5 1 


'fi 


1, 








H 


H 














W 


Si 










o 




- — H- 


— 1-' 


3 




,_4 _ _^ 


CO 




^ 
















1 




35 too 1 






















KJ 










a 


a- 
o 








k! 


.N tt -^ <:n 05 CO >- 


00-400 co.^ 1 




















Ji. 


3 




tstsosstto 


!2i 


3 
















n 








c\ 










<1 


CO 


sggggggggg 


^g 


fi| 




















^ 


gSS&i^feSggSiSiS^ 1 












hj 


O 








M 








t> 


^ 








H 


3 


>-i ^ fc. -J in fcj 


en-ji-' 




■| 


fc^ 


*. ^B iO 00 00 tS X X\ 00 00 








to 




W 


a 








M 








!> 


o 








H 


>^ 


CO ^0 CCH^ 


cofd 


CO 


O 


CO 


Wo ~ o 


5^0 

CD ^- 


3- 






Q, fi. 


a 


O- 


CO 










CO 


5' 








CO 


o. 








to 










3 


O'^r.'^Obj , , 




.5^ 






-! o o 2 P ' • 






> 


5 


P 3 = g 3' 
r: — »- o 7 




o 




tj: -! 2.jK b ' ' 






rV 


re 










B' 


_2 s _^ 


r : J 


1 


tc 


3* 


n'-gs'^ 


' ' ' 




lO 




d::2.cp- 
3 3-3^3 2. ' ' 




» 


o 


: r ; 






p 


re ag.®-? 




re 




c 


re*::; re ^ ^ ^ 




1 








re 




3 J^*© M ' ' 


•- • r- 


'rr 




CO 

TO 

3- 

rt- 


p».*£2 . ' 


:^ :. :: 


CO 

re 






-• re £--^ 








^ 


30®b' 








rt 


O 3 2 o 








pr 


^ ■ § 

K ce 

P P , 

3- 3' 
O O 







Mark. 


; 


5' 2 
re 3 

CO p 


P 


B ^ 

3- 5 2. 


H 

3 

p. 


1^1 


►flGO 
re "O 

si 


ft 

a! 
o 

H 

B 

•2. 


§4 ■ 


Distance, 
center of 

hole to 
edge, ill 

inches. 


O 

as 

o 

l-b 


t 

o 

re 

•d 
c 

3 

o 
3 

CO 

i 


re 
3 
p 


1 

1 






H 

o 
H 



O 

» 

C 
Q 
H 

c! 



s 



H 

fr- 
iz! 
O 

o 



TJl 



Q 
O 
O 

O 

s 



» 

Q 

►• 
H 

I 
50 



374 



NICKEL STEEL FOR BRIDGES 



-"• 



W W V» Cl^ CO CO CC W OI in CO 



^~,\. 5 CO M-\\ "\-C^-S? 

K. Kl\^^^ « 00 l-i .(k CO ►- 



N)t^i^i-'^*^OOCO|*^COCO 



ODOOOfOTCJ'UICOCOCCCOtOtO 

vi'jic^ ooooococoowcn 



30^tJ^tUC7ic;TCOCOCOCOCOK} 

cn-j-joco*><xcncoo-^jcn 



Ml w.^ M w ai -n-i -J oi 

to fO 00 00 to 00 00^^ 00 00 






3 o 2 13 

O 9! — \t- 
■^ ■-! ,, i«N 

S3 M S „. 

£,-» 3 B 

o a> ^<^ 
►1 - o cm 

B B (t ^ . 

«^ob 

tB £, 

P n 

tr a 



.!z| 



P ' 



w 
a 
o 

CO 

H 



Kl 






w 

o 



^ 



C-* >; s^ K, i qj '»! Cq ti o to ^ 



K M OT w^ CO CO CO cn cn cii BO 



- . CO 00 »-* »*^ CO i-i 

COCOCOiO a> K>a 



05 — l-i^COCOt-'^^>-i«NjCJ» 

•-;»o\coa)i-'Oooj*.coco 
OS cots Oi rf^ to iO 



'^OQCtnarOiCJTCOCOCOCObOtCJ 

aioToooooccccoou'Oi 



■^OC4^COib>^COtSCOCO{S6S 

osccoo-jcROocncoocooi 



h-i_i-^u»»-ic;'*ii-^-}c;» 
^-,\,\^^-\^'\ co^^ 

to to 00 00 »0 GO QO^^ 00 CO 






<t 






JOB 



^. § S ,. 



?5 P P 

o-^cn B- 



B P p , , 



ftli 


^t 


m (T> 


a' 


a ft 


°^ 


&« 


"P.S 


?.g 






S2. 

• 0. 


.^^ 


<6 


B & 


l2:n 


(i 


B 


C 


2'< 

i' 




KT' 



Mark. 











B 


"-S 




B-5-s: 




» 


r 




Cfi 


?i 










03 




t: 


5" '2 




P 


D-B 2. 

p. 




B 


B 


m 








0. 






! 








"* 






Tw 












■^ (B 




H 


o2. 






r-& 




, 






> 


S 


S^" 






B 2 d 






r-= p 












►-.ct 

B & 


^2i 




cor? 




B-CB 


oo2 













tJ- 














&p 








ft 2" 






'^ 

















>* 


















cr 













Tl 






01 






■d 


^ 











B" 






fD 


?r 




0. 


n 










c 


















CO 












N 






P 







H 
bd 



b 



Q 

H 



CO 

pa 

G 
Q 
H 

G 

> 



Q 

w 
00 

H 
H 
IS 
C 

> 

O 

O 
> 



00 



O 
Q 
O 
PO 
D 

o 



g 

> 

O 



NICKEL STEEL FOR BRIDGES 



275 



b !»;< c^ K, tl! tr. •*) tq to O to l>^ 



5-H 






^3 



go 



CO M ot -» -' c; o; cc :;ip» or «> 
05 cl H c. ~. S C-. — . c- 3: c; 05 

M.I.S ^- - , 05 X -' *■ CC ■-; 

:^ CO :c to o to c: 



t- CD i-i :p ►- CO h-i -1 -J i-i — 

}0-.\,^^^ 301-^ CO *■.!«>. 

■ S "■ fo " 






OI OI O O O O 36 M CO OJX OI 



0CC&l£k.ij^C;i4^COCCCCC0C0{<0 

cncoMoooQc^oeococn 



— -•-jc;ii-'Cn-»i-'-JV« 

to to X OD to OC C»\ 00 00 



W sow 2W £2 



» S _ -e pas 

2g ao^ 






J5"2 

eg 
"Em 

» ft 



S'- 
2.W 



M TO 






^ ;^ 



a^ =^ 
° o » 

O HI, I 

2 " 2. 

S - o- 

So cs 



crp 

o «■ 





h>!c>,K,l:t:5i'»]t»ll306DV 


n 








5^ 


^*. . 


--.^ 




!Zl H 


CC-c :;» Of 


-■> cc i: CO ii ai a< CO 












— 


— _ — 


- — ~ — ^^* 


1 


1 


























1 n> 




cococoto o> ioa> 


•-' 2. 






bo 5 






1 


«5 o 


««K.-^^..^ 1 


^ -J CO 1-1 










00 CO 00 ^.■^^-v, CO 00 eb to CO oS 


!z! p 






1 










37 








d^ 








!5 *! 


00 00 c;t en en en CO CO CO CO to to 1 










M (B 








r^ 3 














> 9- 








COOIrf^COf..*.COCOCOO-.COtO 










ts -^ 








■l -^ 


^ 


i_j^ 












^5^ 

H 


*. 


to to X 00 to * X-^ X 00 






to 


^ 5- 














c 












c 




2. 
CO 2 


P 

3 


'CD ' 

a. 


P^I'^I^P 


Q. 




p. C" Cb 


Sf ° 








fe 2 








• P 
















g? 


W 


,«« 


.,,,,, ,12! 




'« s 


'''''' 'O 


> 






o 


w 


!^ 


^y^- 






s 

a 




03 




TO 


re 

TO 


' :^ p r : r ;o 

5* 


M 


CtJ 


(T 


r :: r r r :^ :;© 


a:> 





o 


a 

r ' ; :: r r PTO 


M^ 


o 
















a 










ft 







3 






o 


o 


-''--' -CT* 




o 


o 


<D 












cr 


3* 






















o 


C5 






■-J 
















n 


o 






?r 


?r 






m 


w 





Mariv. 





w 










> 




s 


3 


3 


P 








-5 I 



gTOOg.| 

trm (t (s S 
<i- ^s a 

»300« 



t? S 



QQ 



276 



NICKEL STEEL FOR BRIDGES 



* ■ ■ 






35 0505O: 00105 350505 



h^-^l-ii-itSl-n-'fc-^)-' 
CT Ol -Jl-'l-'WtOK>COCT 

►-^ »0 M 05 OS 35 



-os^^-~ 

• -TOSC 

to tots 



05WOW*iOOO;-Q05 



oooo- ootstc 






P P a-S.a.S ^ <i 2 p 
o a >— h-»— ^(t 0)0 o 

>psvfO O O O tl.Z.x»->_i 

if^W\T ~- -. ": ir ^-^^i^„ 

„. _. p p ?j » S [0 _. S- 

B t3 O O - 2 rf -^ 3 • 

J— 'I — cDrD'D^^f:^— o 

2 c o 2.2."^ w K o § 
S 3 p _fB ™ »-j p p= 3 

TO crT 5' S' S- o o ■" J? ■!§ 
ft (C w K P'3'S'cti ft --, 

• ■ ft (t cB '^'5,5' 
/• " • 05 M p S; 



•a-d 
pp" 

a o 

Bp 
Po 

o '•" 

p fi 
1^5^ 



fp 






p o 
m 50 



> 

td 
O 

W 

w 



i— ( 



CO 



§ 



^c:i'-^^SyO_^t)0!3^ 



s » 






5 c?? 

.2. , ,C p j£.D 



uicicjicjrmiyiOTcnc^ioi 

O5O5050505O505C53535 



-^ 00 to '-'^ 00 l-i iO 0^"^ 



c;»*-35cnu'4^cncT5 05c;t 

C-JO03O-^05OC«-7 



OOOCOO- OOQC- ooiots 






^3.Sl (3- 
gjS'a rp 

&2c- §. 



. o 



ooosp'^cO'^ocd 

^"ptc — 3p'"^— 

'^''^pSSi^^pCTOB' 

■-o>-tj?;f5 '^ S-^* i^o 
3 3 2.P~oo^g 

ct> T> p !;o it r'^ tt.tr' 

Qf9 TO ^ 2.- & ^ P TO -t 
aiftBT.;^t-^^TO 

'o'^.fj c ^■^■•'■- ~'2.* 

PP So rf'o EL P £. 
ft CD i^ ■; o K..': eta 

o o 5 <t • g &"© g 



o o 



p p -• p 

S- B- oc - 



TO 






S.S.M.O. .01 —; 



w 



ii 



Original, 
in inches. 



Enlarged, 
in inches. 






-2 ' 



_ ) go 
ItooSI' 

p-tt ct CD P 

t" ^-- a 



o 



o 


^ 


w 


o 


;> 


W 


o 








!2i 


Hi 



NICKEL STEEL FOR BRIDGES 



277 



TjC^'S.iliC^^jC!?;^ I O 









s! S ™ s: 



at UT CJi in c o» :p c;» ai crt 



l-'i-'tOi-'i-'tOtOi-'i-'.-i 
^^ i-i CO CC CC CO •-- v-i 



o»a>-itoo50ocoooc 
-locooe'^oew-^c 



qcodod'- 0000- ootsts 



^- - =* ^ ■- w^ »2 
— - "S ■ -J . o. 



fo-_lT_ — 'DC": 
i^^r- jT — a n * ^~ 

— _.o aJi; t ^ P 

ts t3 ?ro ^ c ;; 2.E s 

• '■" 2 " 5 P M ' — 

»2 2* 2. pit 5^ J 2 

Vb"^ rr —^ .^ tr 5" 

2B'°' oii 5 2 



» '4 tt> 

R 9> CC 

?ro. 



o 1 

H 

E » 

D - 



P o 
— SO 



k! 

M 

d 



o 



§ 

C5 



■^cjS^Sychjpi?:;; 



5)33 

H 

CD 2 



J. D m 2. B 

"> S ? 5 5f 

CL CD 2 O.'^ 



o; -J -I -) 00 oc 

O M cc cc *^ o 



M-J WC 



00 00 00- ocoD- OS IS to 






KH^ CO 

■ — <D — r 50 

o-SP- 2. 



■1 T ~o S" 

so T - c -! 

o P (t T a 

?r ?r (T j^, CD 

3 g !.►:• 



2 » n -, - 

Oc?;pP 

B fS-!;- 

(t o —B^v-v 

0.(t S Op 

Oq -. — ^ -^ 

_ ns . _ 

2-5:?; 2,g 
2- 1£ £■ fS* 2. 



^? c 2 S 

(T " ^ p o 

•/i -5 3 -i-.\f 

str^ p TQf^ 
"■pS,o 



.. .- J 5 ST 

c_ „ ;.; ct ^ 
55,5.05 



i^.^2' 



O! 



SO ffg 



foil 

« 5.? ' 

CO 



■n 




P 




















^ 


^ 


B 






?7 


C 







■1 


P 


^ 
P 


P 


»■ 


W 


o5 




0! 




» 


P 







cf- 







1: 



a 




>■ 
W 
W 

b 



ss 



Original, 
in inches. 



Enlarged, 
In inches. 



® gB 



B"CD tc n> p 






M 



1-3 

d 

o 

d 
tri 



B 
QO 

O 

o 

w 

ca 
O 
!zi 

CO 
H 
W 
W 



o 

CO 

►^ 
O 



278 



NICKEL STEEL FOR BRIDGES 



xjc)is?5iw:)_*^pta.fe 









OOOStOfOl-'l-'l-'l-'CRO 



OOOOOrj- oood' 00»0lO 









n f' '* 2 ~ 2 ^ -? o o 

OOSSSO-ioJmOO 
S B 2.^ 5^95 o o B d 

3 B^? £.:i3-c-3 3 
<t> * %-t' ^ S 2.2.fe 2 



c*5!-tt 

(5' cc ■• • 


W K 


— 


S'C rt-rt- 


O 


'D " B O O 
02 K-2.B"&' 


-• o 


B 




C/3 '^■" 


ag 
O B f" 


O 




R * B S 

" S 2 3 
S ^ f^ S 


pl 

95 W 3 


'I 
95 
O 




C 
B 




eg w ^ 

.• 55 p » 


Vr*X 


rr 




c n cr? 


O B-' 


2 




0^2.2. 
g.§2i^ 




CD 




2.0= K^ 








(I a. rji 


B m 

B-fT 






2. 2 



^c:;'^?!tcoh30!s(^ 



C5C502C3C5asC5CiC^OS 



^"^J* :rt '~'J=o ^ en CT ► 



— "-'■^ \ i-"^^ QC OC ^ i-> 

C; 05 — •-* 05 '-^ COOi 






ODOo' QOiStS 



o ™ o 



;p —ft ,E2 2. ^ 
S 2 S' ' 5" o' ?" 



9=<T>*SCn;SH?95p 
' B B 



_. o :i 95 '^ 2, o : 

T o "= ^ n " • 



crs 






?5'«W 



W 



3 sp-B^tJ B-g B 

2(rJsa5-;r5T2 2 

-^^' - w ^ B'« W "? 
'''^95'C^B've'^'* 
O ° _,S B 2.cD o o 
B-Og^l&^'^^g-ffB- 

2.B-;5-B-.^^ ,^oo 

B m B"' - B-g ^O O 
;^g.g-po2.C!gB-o„ 

P O o 95 !» o t^* P 9» 

o_,-2 X n SfacLg^g 

Brera.2.7> 2- 
£: o ,. " » a !?• 
«'='i-2." 2» 

M W (I C- 

SS- 2 



Original, 
in inches. 



Enlarged, 
in inches. 



17 H 

?3B 

n5e 



B oagS- 

» » t6 >1 g 





r/3 




H 




W 




d 




n 




h^ 




ci 




» 


o 


>- 


n 


t2j 


w 


n 


> 
o 




!2! 


03 



o 

K 

o 

H 
S 



NICKEL STEEL FOR BRIDGES 



279 



TABLE 45. — "Braring-on-Pins'' Tests: To Determine Amount of 
Yielding of Pin and Supports. 

Readings taken with Olsen's Compression Micrometer. 









Micrometer Readings, in 


Inches 








Load. 


Pin No. 1. 


Pin No. 2. 




No. 1. 


No. 2. 


No. 3. 


No. 4. 


No 5. 


No. 6. 


No. 7. 


No. 8. 


No. 9. 


2 ono 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


*5 000 


0.0018 


0.0014 


0.0009 


0.0011 


0.0017 


0.0014 


0.0016 


0.0008 


0.0011 


10 uoo 


0.0035 


0.0023 


0.0019 


0.0022 


0.0031 


0.0026 


0.0029 


0.0017 


0.0028 


*15 000 


0.0047 


0.0033 


0.0027 


0.0029 


0.0037 


0.0030 


0.0034 


0.0022 


0.00351 


20 000 


0.0053 


0.0040 


0.0034 


0.00.36 


0.0042 


0.0033 


0.0039 


0.0025 


0.0040 


*25 000 


0.00f)4 


0.0017 


0.0037 


0.0042 


0.0048 


0.0039 


0.0044 


0.0026 


0.0042 


SO 000 


0.0072 


0.0053 


0.0O40 


0.0018 


0.0054 


0.0044 


0.0049 


0.0029 


0.0048 


*35 000 


0.0077 


0.0058 


0.0045 


0.0050 


0.0059 


0.0017 


0.0051 


0.0032 


0052 


40 (K.0 


0.0081 


0.0063 


0.0O47 


0.0053 


0.0065 


0.0050 


0.0054 


0.0035 


0.0054 


*45 000 


0.008« 


0.0068 


0.0050 


0.0054 


0.0070 


0.0052 


0.0056 


0.0038 


0.0057 


50 000 


0.0089 


O.0O73 


0.0053 


0.0058 


0.0074 


0.0056 


0.0060 


0.0041 


0.0081 


*55 000 


0.0092 


0.0077 


0.0056 


0.0001 


0.0076 


0.0059 


0.0061 


0.0012 


0063 


(10 000 


0.009.1 


0.0081 


0.0061 


0.0064 


0.0080 


0.0062 


0.0063 


0.0045 


0.0066 


*(;5 (X)0 


0.0099 


O.OCH 


0.0061 


0.0065 


0.0083 


0.0066 


0.0065 


0.0047 


0.0068 


70 OOO 


0.0103 


0.0088 


0.0066 


0.0067 


0.0087 


0.0070 


0.006K 


0.0049 


0071 


*75 000 


0.0106 


0.0093 


0.0068 


0.0088 


0.0097 


0.0073 


0.0070 


0.0952 


0.0075 


80 000 


0.0111 


0.0C96 


0.0071 


0.0071 


0.0105 


0.0075 


0.0072 


0.0053 


0.0078 


*85 000 


0.0116 


0.0101 


0.0075 


0.0074 


0.0110 










90 000 


0.0121 


O.OIOC 


0.0079 


0.0076 


0.0118 










*95 000 


0.0125 


0.0110 


0.0083 


0.0078 












100 000 





0.0117 


0.0089 


0.0081 




1 









Note.— This table is condensed from the original. 
♦Interpolated values in horizontal lines marked in this way. 



280 



NICKEL STEEL FOR BRIDGES 



TABLE 46. — "Bearing-on-Pins" Tests: Structural Plate Materul. 

Readings taken with Olsen's Compression Micrometer. 

Diameter of Pin = 1 in. Length of Piece above Pin = 3.80 in. 







Micrometer Readings, in Inches. 




Load. 














Nickel Steel— Heat Ko. 17 673. 


Carbon Steel— Heat No. 33 342. 


2 000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


*5 000 


0.0017 


0.0010 


0.0010 


0.0011 


0.0013 


0.0017 


10 000 


0.0030 


0.0018 


0.0020 


0.0025 


0.0023 


0.0028 


*]5 000 


0.0041 


0.0027 


0.0031 


0.0035 


0.0033 


0.0041 


20 000 


0.0050 


0.0037 


0.O040 


0.0045 


0.0042 


0.0052 


*25 000 


0.0057 


0.0045 


0048 


0.0055 


0.0052 


0.0063 


30 000 


0.0064 


0.0053 


0.0054 


0.0067 


0066 


0.0076 


*35 000 


0.0072 


0.0060 


0.0060 


0.0083 


0.0086 


0.0098 


40 000 


0.0082 


0.0007 


0.0067 


0.0111 


0.0118 


0.0124 


*45 000 


0.0091 


0077 


0.0073 


0.0143 


0.0156 


0.0168 


50 000 


0.0101 


0.0087 


o.ooaB 


0.0193 


0.0206 


0.0206 


•55 000 


0.0112 


0.0100 


0.0095 


0.0255 


0.0271 


0.0265 


60 000 


0.0122 


0.0120 


0.0108 


0.0333 


0.0359 


0346 


*65 000 


0.0135 


0.0129 


0.0121 


0.0432 


0477 


0.0450 


70 000 


0.0148 


0.0142 


0.0135 


0.0551 


0.0625 


0.0.563 


*75 (X)0 


0.0163 


0.0151 


0.0148 


0.0696 


0.0803 


0.0693 


80 000 


0.0182 


0.0167 


0.0107 


0.0850 


0.0997 


0.0871 


*85 000 


0.0204 


0.0188 


0.0188 


0.1031 


0.1201 


0.1074 


<)0 TOO 


0.0229 


0.0212 


0.0210 


0.1212 


0.1420 


0.1290 


*95 000 


0.0256 


0.0238 


0.0236 








100 000 


0.0288 


0.0263 


0.02S5 









Note.— This table is condensed from the original. 

* Interpolated values in horizontal lines marked in this way. 



NICKEL STEEL FOR BRIDGES 



281 



TAJjLE 47. — "Bearinq-on-Pins" Tests: Structural Plate Material. 

Amount of Compression in Test Pieces. 

Corrected for Yiehliiig of Pin and Supports. 

Diameter of Pin = 1 in. Length of Piece above Pin = 3.80 in. 







Amount of Compression, in Inches. 




Load. 














NicJcel Steel— Heat No. 17 673. 


Carbon Steel— Heat N 


o. 3.3 342. 


2 000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


O.OOOO 


*5oa) 


0.0004 


-0.0002 


-0.0001 


—0.0003 


—0.0001 


0.0002 


10 000 


0.0007 


— O.U004 


—0.0003 


0.0002 


0.0000 


0.0')05 


*15 000 


0.0009 


—0.0004 


0.0001 


0.0002 


0.0000 


0.0008 


*20 000 


0.0011 


—0.0001 


0.0003 


0.0005 


0.0002 


0.0012 


*25 000 


0.0011 


0.000 1 


0.0006 


0.0008 


0.0005 


0.0010 


30 000 


0.0013 


0.0003 


0.0007 


0014 


0.0013 


0.0023 


*35 000 


0016 


0.0007 


0.0010 


0.0025 


0.0028 


0.0040 


40 0(j0 


0.0022 


0010 


0.0013 


0.0048 


0.0055 


0.0061 


*45 000 


0.0026 


0.0017 


0.0017 


0.0075 


0.0088 


0.0089 


50 0C0 


0.0032 


0.0022 


0.0022 


0.0120 


0.0133 


0.0133 


♦55 000 


0.0039 


0032 


0.0031 


0.0178 


0.0194 


0.0188 


fiOOOO 


, 0.0045 


0.0048 


0041 


0.0252 


0.0278 


0.0265 


*65 000 


0.0O55 


0055 


0.0051 


0.0348 


0.0393 


0.0366 


70 000 


0.0065 


0.0064 


0.0062 


0.0463 


0.0.5.37 


0.047'5 


*r5ooo 


0.0076 


0.0070 


0.0072 


0.0603 


0.0710 


0.(J600 


80 000 


0.0092 


0.0083 


0.0089 


0.0754 


0.0901 


0.0175 


*85 000 


0.0109 


O.OIOO 


0.0106 


0.0930 


0.1100 


0.0973 


90 000 


0.0130 


0.0120 


0.0125 


0.1106 


0.1314 


0.1184 


♦95 000 


0.0153 


0.0142 


0.0147 








100 000 


I 0.0179 


0.0102 


0.0172 









Note.— This table is condensed from the original. 

* Interpolated values in horizontal lines marked in this way. 



282 



NICKEL STEEL FOR BRIDGES 



TABLE 48. — "Bearing-on-Pins" Tests: Plate Kivet Material. 

Readings taken with Olsen's Compression Micrometer. 

Diameter of Pin = 1 in. Length of Piece above Pin = 3.80 in. 



• 




Micrometer Readings, in Inches. 




Load. 














Nickel steel— Heat No. 2 096. 


Carbon steel— Heat No. 19 241. 


2 000 


0.0000 


0.0000 


0.0000 


0.0000 


O.OUOO 


0.0000 


*5 000 


O.OOIS 


0.0011 


0.0014 


0.0017 


0.0011 


0.0015 


10 000 


0.0034 


0.0021 


0(133 


0.0033 


0.0024 


0.0029 


*15 000 


0.0044 


0.0031 


0.(1(145 


0.0048 


0.0037 


0.0044 


20 000 . 


0.0053 


0.0039 


().(M).">4 


0.(1064 


0.0053 


0.0061 


*25 000 


0.00(10 


0.0046 


0.(i()(i.3 


0.0079 


0.0068 


0.0079 


SO 000 


O.OOCG 


0054 


0.(J072 


0.0115 


0.0100 


0.0108 


*35 000 


0.0075 


0.0063 


0.0083 


0.0160 


0.0143 


0.0148 


40 000 


0.0083 


0.0072 


0.0096 


0.0206 


0.0190 


0.0194 


*45 000 


0.0197 


0.f»079 


0.0115 


0.0266 


0.0247 


0.0338 


50 000 


0.0114 


0.0092 


0.0132 


0.0340 


0.0313 


0.0290 


*55 000 


0.0135 


0.0107 


0.0151 


0.0482 


0.0431 


0.f)364 


60 000 


0.0158 


0.0126 


0.0177 


0.0635 


0.0568 


0.0487 


*65 000 


0.0181 


0.0150 


0.0302 


0.0800 


0.0732 


0.0612 


70 000 


0.0205 


0.0179 


0.02.31 


0.1003 


0.0885 


0.0785 


*75 UOO 


0.0234 


0.0217 


O.0-.i64 








80 000 1 


0.0371 


0.0259 


0.0298 








*85 ono 


0.0308 












90 000 


0.0.359 














Note — This table is conciensed from the original. 

* Interpolated values in horizontal lines marked in this way. 



TABLE 49. — "Bearing-on-Pins"''' Tests: Plate Eivet Material. 
Amount of Compression in Test Pieces. 

Corrected for Yielding on Pin and Supports. 
Diameter of Pin = 1 in. Length of Piece above Pin = 3.80 in. 



Load. 


Amount of Compression, in Inches. 
















Nickel Steel — Heat No. 2 096. 


Carbon Steel — Heat No. 19 241. 


2 000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


* 5 000 


0.0003 


0.0000 


0.0005 


0.0002 


—0.0002 


0.0005 


10 000 


0.0006 


0.0001 


0.0013 


0.0005 


-0.0001 


0.0005 


*15 000 


0.0011 


0.0006 


0.0019 


0.0016 


0.0009 


0.0013 


20 000 


0.0017 


0.00)0 


0.0024 


0.0038 


0.0029 


0.0t>26 


*25 000 


0.0018 


0.0015 


0.0031 


0.00.37 


0.0031 


0.0042 


80 000 


0.0019 


0.0019 


0.00.37 


0.0068 


0.00.50 


0.00(57 


*35 000 


0.0025 


0.0025 


0.0044 


0.0110 


0.0098 


0.0103 


40 000 


0.0031 


0.0031 


0.0054 


0.0154 


0.0143 


0.0146 


*45 000 


0.0042 


0.00.35 


0.0069 


0.0211 


0.0197 


0.0187 


50 000 


0.0056 


0.0045 


0.0084 


0.0282 


0.O2G0 


0.0336 


*55 000 


0.0075 


0.00.59 


0.0104 


0.0423 


0.0377 


0.0307 


()0 000 


0.0095 


0.0077 


0.0125 


0.0.572 


0.0.511 


0.0428 


*65 000 


0.0115 


0.0097 


0.0148 


0.0734 


0.0673 


0.05.50 


70 000 


0.0136 


0.0124 


0.0174 


0.0934 


0.0823 


0.C721 


*75 000 


0.0162 


0.01.59 


0.0304 








80 000 


0.0197 


0.0199 


0.0236 









Note.— This table is condensed from the original. 

* Interpolated values on horizontal lines marked in this way. 



NICKEL STEEL FOR BRIDGES 



283 



1 3. 



SI 
o 

w 

H 
> 

H 

w 
o 

< 



> 



<1 

t> 
fl 

d 

w 

on 
O 

c 
t-< 

W 

t> 

t?^ 

H 
O 

O 

td 
o 

!2i 

QO 
^^ 

t?i 



22. S"? 
So 2.2 



g28g 

CO -5 1:^ OS 



pppp 
o'oo'p 



> 



pppp 
o'o'o'q 



w 


w 


*t3 


►x) 






















«1 




p 


85 










(t 


(t 


% 








a 




t3 


o 










93 




-3 

c 

o 










tJ 












to 




i-t 






. 




o 


,_J 




5^ 




p 




s 


to 


P 










^>_ 


.-^1^ 


.-^:^_ 


-J^ 




















H 










^. 




C 


ooo 
'■<> -^ -J 


otoo 


ooo 


o o o 




B 


o 




cnaioi 






^ 




ooo 


ooo 


ooo 


























•a 














» B 


_t?=) 


CTOsCn 


tcce w 




co«. >c>. 










c: O* 


ODOCO 


S'-x X 


3? 


oioto 








■ p 




888 


^§d 


o o- 


888 


B 


^^'b- 










f6 


w 




ooo 


0005 


oo- 




-H 




s 
















ooo 


oo- 




3 » 

5 CO 




2m§ 
















■ a 






00*^*. 


CO^IO 




0.<^H- 




a 














o *^ 


ooo 


ooo 


ooo 


ooo 


g 




III 


ooo 


8SS 


ooo 




>z. 










•8 B 

11 












OcDO 


OOti^tCk, 


>C^i£09 


-^ t0O5 




























ooo 


ooo 


ooo 


ooo 


a 




SoS 


ssg 




goo 












QO WCD 


CS o o 




4^ O Ci 


o 
















:a 












ooo 


ooo 


ooo 


ooo 


g 




s« 


ooo 


o o *-* 


















o5 


SJ3S 














*:*:~j 


oxx 


° 




S.S 
d 












ooo 




ooo 


ooo 


^ 




a 


ooo 








or 




























^ 




w 


o*. *. 




OOT*. 


o o: fO 




IStO^ 


ooo 


8 8 8 




bO 




O 




oc o 


M "^ 




g 


, ^ ^ 


, ., ., 


-- r r 


.. -C 


BJ5 




E4 








r ..^ 














(t so 




CD 










ggs 


oco 


OOO 


ooo 


wi^ 




CD 
























O 




















SP 




1=9 


K^ ^ . 


. . . 


. . . 


.*"* • 1 


poo 










5 


-.coo 




ONJO 


























. - c 


" o 




a 






















5"^ 




f 

H 






cob 


b b b 


JO o. 




COfOI-' 


C0i4^C0 












0.01 
0.01 
0.01 


o o o 
p o b 


ooo 

§28 


ooo 

83'S 


b'"ci 
B-5'c 


ft B SS 



H 


O 




5- 
o 


O' 


H 








lA 


'_^ 


c 


n 


."^ 


td 


a- 


C 


tr^ 


C+. 


w 








ai 


(U 


c» 


O 


•— ' 


o 


l-h 






po 


g 


1 


13 


P5 


;; 



D 



^ §? 



?8 g 

o d 

« 2 



l-W cr 



td 
> 

M 

o 
6 

!2! 



a 




rf 


(yj 


"^ 


H 


O 




03 


!2{ 



o 
W 

CO 
W 

o 

o 

K 

o 









O 



284 



NICKEL STEEL FOR BRIDGES 



o 

o 
» 

5' 

OR! 




50 


g 


sesoo 

53 700 

54 600 

56 400 

57 300 

58 500 

60 500 

61900 




Vt -u 




7: 






s 


yr 






\ pppppppppppp 

o o a- ac CO M iiT oi oc oi w X 


• o o o o o 

■ COCOiOiCiC 

• cnoxaio 


- 










































o> 
















p 


p 

i 


o 

i 


ppppp- p 

C/T OI O" .&. *. • CO 

c«ceoc«co; cn 






''3 oooooooooooo 
=: g 2 2 2 S g S g S b 2 2 










1 






|8S 






































*1 

s. 












o 

c 

CO 


o 
o 


i 


oooo 

ggg's 

J^4-CO — 




























N) 










































M 










































>«>■ 











































































































































































NICKEL STEEL FOR BRIDGES 



285 







e. ^ ^ 


??1 


iSS^if.^ 


f; 


-' = =;' = 


?S f s « « tc to 1 
S 'i Z- Z = " ' 


C :r X -i- ii 


: 




ft D- 

' Tl 
(0 

1 


1 


oppj 




: si 


; gS: 






: : o^ : : : 


;8M 


P !ZI 
o p 


5 

11 

(C o 

(D 

13 

3' 


o 
c 

CO 
H 

r 




















: !z! 

: P 

; to 


IS: i: : 


co' • 










lih 


5*; 


p 

JO 




















p 

"o 

§ 


p 


3 

s 
1 

5' 

5' 
o 




















o 


p 

6S 




















p 


p 

OS 








I: ^ 


5j P: : 






o^ : : : : 

i! : : ; • 






!2! 

p 


B 
■a 
o 

pS 

ft m 

p-tr 

C6 O 
OJ -J 

5' 


IT" 


; 








P: : ? 

lb.'- '• H 

O; ; C 


2; 


: : o^ 


pj j 1 po 


: g: : = 


p 


p 














;: COOCO 
o; cnoutoc 


^. . . 


p 
o 


p 

CO 








= £.« 


s| P: : 






P: : : : : 




o 


!2; 

p 


3 

§ 

1 

B' 
S' 

o 

S' 
ft) 

CO 






: : "^i : : 

. . p. . . 

• • (T- • • 
: : a: : : 


?: : .'^ 


0^ 


■ '-^ ■ 


b: : : : : 














p : : 

a"- '■ 
p.; ; 




o 


'g'2: = 

MJC; < 


2i 8SSi i 

^ • 50 X CO • • 




p 


p 

CO 










o 




o 




s 




•fl 


g 


oa 


JO 


00 


a. 

CD 


1 


o 




p 


H 


to 


00 


l-l 


H 


05 


00 


O 


o 


O 


"a 


O 

o 


OQ 


a" 




^? 


1-3 


&< 


M 


"ri 


O 


P 




n 


•^ 




H 






o 




H 


^ 


o 


t?! 




a 






S 




OQ 


Ul 


g 




O 


CJ 


cr 


H 




C-! 


p 


^ 


rt> 


> 


■^ 


C 


9 




p 


N 




t4 


X 


c 


M 


CQ 


ri 


H 


P 




o 


fi 


o 


> 


H! 


!z! 


"rJ 


a 


S 


^ 


"-<! 


M 




O 








Q 




K 




O 




> 








S 




!z! 




OQ 




i;^ 




M 




tr" 



286 



NICKEL STEEL FOR BRIDGES 



1 —Cambe 

Compr 

ther 

2— Cambe 

3— Cambe 


ff 


- 


S'2 


1 

c 

S 

'a" 


-! '-i ft ft -^ 


H O 




ap- 


P ::so 


failure, 2. 
meter wa 
e, are omi 
no load, 

13 900 lb. 

failure, 2 
no load, 

22 200 lb. 

failure, 


» 






p -p. 


N) 


►-=^' 


'n^B-^^S^^ 


y 


g?j^ 


B-:?B-2?^tp 


s 


B^d 


down 
odged t 
from til 

up. 

q. in., 
.up. 

up. 

q. in.. 
. down. 


o 

t3 


Is? 




• o 


U w ^^ ^ 




o 














B B a'fB 




a 



O 

•-3 

a 
1 

S' 


38 000 

38 900 

39 800 

41600 

43 500 

46 300 


*^ 


: : : : : : poop 


Oi 


'■ '.'.'. ooooop 
; '.'.'. ■tncji'*-'*-')!!-'*- 

: : : : gSgSgg^^S 


IS 


0.403 
0.415 
0.420 
0.426 
0.428 




• poooooop 




0.001 
0.005 
0.005 
0.008 
0.009 
j Failed. 1 
I 0.010 ( 




: : : : -^ : : : : 

: ■. : : B.P'. '. '■ '■ 
: : : : ctS' '■ '■ '■ 

■ ■ ■ p-O; ; ; ; 


- 




OS 




Ot 



















o 



Si. 



NICKEL STEEL FOR BRIDGES 



287 





u 


CO 


CO 


cocc 


CO 


CO 


iC 


JO 

■s 




to 
2 


5 S 


tc 




g 


ff 


K) 


t? 








; M 


ii 




2S 


JO 






II 




0.435 
0.445 
0.455 
0.470 
0.490 


o 


: P 


• p 






: ."^ 








• >* 

: ° 




000' 

CnSo; 




2 

p 


H 

CD 

3 

-■§ 

5-3 

CD 

5" 


M 


eg 


0.388 
0.395 
0.415 
0.433 
0.430 
0.433 
0.448 


o 


o 

s 

CO 


pppp 
coiojoVo 

wxx3o 
















CO 








ooooooooocooccooooooo 

OOOMCnComOltnOUlODOOCTQOCBMOOOaiOO 












i 


p 

CO 


0.001 
0.001 
0.001 
0.001 
0.001 


o 


o 

s 
































p 


p 


3 
;o 

D 
(D 

p 

rt- 

1 

5' 

5' 

























































g 


p 




















































p 


p 

CO 


















































p pppppppp 
CO io V« 'fo io is Vs Vo Vo 

C-.0 3C«-3=:4iJ.CO 

cocococcoocacoo 




pppppp 

OCOOCQCX 30 


p 


1^ 


3 

J 

g-g- 

1 

d 
55 


Q 

i 














o 


o 

i 


p 

CO 


o 

CO 


pj p 
i: 1 


p p p p 
McctS — SS^SE 


p 


1 

JO 




















ppppp ppp pp 
'*. i. CO o; CO w CO Vo ioVo 

CO-i'XCt.».»OOOC-505 


p 
ts 

2 


c 


; P: 
: 2: 




p 


p 

CO 


















a"- 



































p 


3* 

p 

• 1 












^^■- 

a> iff 

COS; 


p': 


o- 

ii 


o' 

ii 


o- 


0' 




OOC:OC:COOO- 

SSSgg'gg'SS: 

ccoc:-j;c<TCccc — >-| 


p 


p 

JO 


















j 

f 

j 


5 


ii 


5 5 

DC 

5c 


= 05 

DOC 


= < 

.nc 


DC 

d' 


d; 

1^ 


¥ 




D- 

2: 




D 
D 

g 


c= 










8 




!Z1 

p 

CO 



o 










g 











>t3 







rn 


a 




<n 


t<j 







i-i 




!z! 


Ci 





H 









H 


"J^ 




{/J 








y 


's-' 




CO 


w 


1— 1 


v; 





P- 






(-1 


5 


1 

0' 














<n- 





T1 


S 


!^ 


H 




GO 
H 





P 


H 


!z; 




tri 


uj 




















a 




r/3 


^ 






^ 




r^ 


CI 







k! 




a 


1— 1 




> 



1^ 



w 

CO 
« 
It" 



388 



NICKEL 8TEEL FOR BRIDGES 





1 . 




















II 




05 


(X 


00 


00 


0-. 


OS 


OS 


OS 






r 


' 


r. 


r 


p 


7 


:; 


^ 


Size, in inclies. 




K> 


?* 


tB 


to 


r 


r 


^ 


r 






CO 


s 


3 


CO 


i(^ 


>f>- 


K^ 


»^ 


Width, 


l_rO 


X 




s 


OS 


CB 












in inclies. 


2,gii 


IS 












CD 


O 






c j5 


D 
> 




ll^ 


-J 


-» 


-» 


o» 




cn 


en 


Diameter, 




o 




O 


o 




cn 


cn 


cn 


in inches. 










K) 












1*^ 






IS 

5 




OS 


-» 


00 


-» 


OS 


OS 


OS 


OS 


Elongated. Diam- 




■^ 


on 


o 


00 


■^^ 


o 


to 


to 


eter, in inches. 


K 










o 






o 


OS 




r 

H 




Ol 










OS 












o 


-a 


o 


cn 


o 


CO 


OS 


Excess, per cent. 




l-l 


03 


Ul 


00 


OS 


o: 


-J 


-3 








ss 


to 


CO 


CO 


*. 




•u 


i(^ 


Width, 


E..2i ' ^ 






cn 














m inches. 






to 


10 


to 


o 


to 








adB 

gina 
men 
ons. 
























1^ 


•<l 


-.1 


-» 


en 


cn 


cn 


cn 


Diameter, 




o 


o 


c 


o 


2 


cn 


cn 


cn 


in inches. 


^1 1 II 






















>V 




OS 


-J 


-J 


-» 


OS 


OS 


OS 


OS 


Elongated. Diam- 


S 




CT 


00 


<! 


00 


o 




;_! 


CO 


eter, in inches, i CC 1 1 
















































o 


00 






-* 


■Ci. 


OS 


CO 


Excess, per cent. 






o 


05 


t-i 


tn 


o 


CD 


o 


o 








is 


































to 




Before fracture. 




>l^ 


^ 


CO 


ts 


-J 


-:? 


<! 


-1 




• ai4 ? 








50 
















ts 


i§ 


10 


to 




to 


to 


to 














t+^ 




cn 


After fracture. 


S" "il (B ^ 


¥; 
^ 


00 


g 


g 


fe 


3 


s 


oo 


li; 




,_l 


1.^ 


,_t 








g 






p> 


1(^ 


CD 


CO 


CO 


o 


O 


o 


Gauge length, in feet. 




o 


O 


o 


o 


o 


o 


o 


b 




o 


CO 








cn 


en 




cn 


Original Dimensions. 


^ 






■^ 


o 




CD 


CO 


CO 


Area, in square inches. 


^ 






M^ 






OS 


























►V 


.'^ 


OS 


OS 


pj 


ho 


cn 


en 

b 


^ 


P CO 


l?o 




o 


o 


00 


CO 


o 








B a 




(t> 












^ 




o" 


>■ s 




^ 


'>u 


OS 


cn 


p 

OS 


p 

OS 


p 
■-J 


p 




a CO 
So 






>«>' 


w 




OS 












ts 

en 


o 


li 


o 


CO 


o: 


CO 


CO 


Area, in 




-^ 


§ 


•(>. 


o 


-* 


cn 

tn 


;^ 


-cn 


square inches. 






« 


2 


cn 


•Ci. 


cn 


cn 


cn 


cn 




^^ 




|C>. 


to 


CO 




ic 


,c>. 




Yield point. 


Load 

OUNDS 
UARE 




8 


^ 


o 


o 
o 


o 
o 


o 
o 




o 














l_l 


1^ 


^^ 








S 


s 


s 


dB 


« 


a 


o 


s 


Ultimate 


5! 2 2 




~Ol~ 


i 


00 

8 


to 


CO 

s 


to 

8 


1 


o 

s 


strength. 


O B ^ 




UT 


Vt 


OI 


cn 


cn 


cn 


V. 






J^ 
















Elastic ratio, per cent. 




o 


!S 


OS 


CO 


CO 


-1 


OS 


CO 






B 


g 


^ 


iS 


^ 


to 


8 


s 


12 inches. 






1^ 


00 


00 


to 


00 


to 


CO 


to 










»-A 
















■o- 


1^ 


to 


to 


o 


o 


'-' 


to 


10 feet. 


3n M S 




cn 


to 


o 


o 


OS 


^ 


CO 


cn 




"s^i 








k-1 t-l. 












coit^tci 


»-i 


O tD!0 


CO 


CO 


t-t. 


20 feet. 


o "? 




05??CH 


CO 


cnS^w 


OS 


cn 


o 


»! 




l§ 


lb. 


£5 


^ 


OS 


g 


£S 


fe 


Reduction of area, per cent. 




00 


00 


•(>- 


o 


>-*■ 




00 


OS 












4t 


rn 


rn 














d 


73 


J/J 

1 






CO 

13 




Character. 








•p 


•d 


0<5 


V 


V 


(Jl) 


(jt) 




a 
w 






00 


CO 


S 


CO 


4^ 


■^ 


CO 


Head A. 




'to 


a 


to 






OS 


00 


■^ 


Location, 


m feet. 











td 

> 
93 



w 

H 
H 
tr- 

td 

> 



NICKEL STEEL FOR BRIDGES 



289 



TABLE 54. — Tensile Tests of 6 by 1-in. Eye-Bars^ up to the Yield 
Point. Eye-Bar Material — Nickel Steel — Stretch and Per- 
manent Set. 



Load on 
bar, in 
pounds 

per 
square 
inch. 


B 1. Area = 5.96 

IN. 


B 2, Area = 5.96 

IN. 


B 3, Area = 5.96 

IN. 


Load on 
bar, in 
pounds 

per 

square 

inch 


B4, Area = 5.88 

Of. 


Stretch. Set. 


Stretch. 


Set. 


Stretch. 


Set. 


Stretch. 


Set. 


■8466 
16 800 
25 200 
33 600 
41900 
43 600 
45 300 
46100 
47 000 

47 800 

48 700 

49 500 

50 300 

51 200 

52 000 

52 900 

53 700 
.54 500 
55 400 

5t' 200 
.57 100 

57 900 

58 700 

59 600 

60 400 
67 100 


0.000 
0.047 
0.100 

0.222 

0.366 

6.3i3 
0.313 
0.322 
325 
0.388 
341 
0.347 
0.3.56 
0.363 
0.369 
0.375 

0..388 
0.394 
0.406 
0.413 
peeling 
0.431 


0.000 
0.000 
0.000 

6.'666 
b'.m 

0.005 
0.006 
0.008 
0.009 
0.009 
0.009 
0.009 
0.009 
0.009 
0.009 

0.011 
0.013 
0.016 
0.019 

0.038 
0.063 


0.000 
0.063 
0.144 
0.175 
0.231 
O.300 
0..303 
0.325 
0.328 
0.33H 
0.350 
0.363 
0.363 
0.369 
0.375 
394 
0.400 
0.406 
0.406 
0.419 

0.428 
6.'456 


0.000 
0.000 
0.000 
0.000 
0.000 
0.006 
0.000 
0.000 
0.006 
0.006 
0.013 
0.019 
0.022 
0.023 
0.023 
0.025 
0.027 
002» 
0.031 
0.034 

0.041 
6.056 


0.000 


Q.iii 

6.325 
0.331 
0.3.38 
0.3.50 
0.363 
0.363 
0.375 
0.375 
0,388 
0.394 

(6.4191 
0.444 

6.'456 


0.000 
0.000 
0.000 

0.666 
6.'{)66 

6.006 
0.009 
0.016 
0.035 
0.028 
0.031 
0.034 
0.038 
0.041 
0.044 
0.050 

0.056 


6081 
6.506 


'8 566 
17 000 

34666 

45966 

47 666 

48 500 

49 300 

50 200 
51000 
51900 

52 700 

53 600 

54 400 

55 300 
56100 

57 000 

59566 
60 400 


0.000 
6.094 

olaoe 

6.'288 

6.303 
0.313 
0.319 
0.325 
0.338 
338 
0.350 
0.363 
0.369 
0.381 
0.394 

0.406 
peeling 


0.000 
0.000 
0.000 

6.'d66 

6.'(id6 

o.'dog 
0.011 

0.013 
0.013 
0.014 
0.016 
0.019 
0.020 

6."623 
0.027 

0.038 
6.'338 


Ultimate 
strength 
Order of 
testing 


99 000 
4 


1U1500 
1 


103 200 
2 


100 300 
3 



The stretch and set were measured in a length of 15 ft.; they are expressed in 
decimals of an inch. 

The first readings for Bars H 3 and B 4 were discarded because the pointer of the 
extensometer did not assume a peraianent zero until after a load of 100 000 lb. had been 
applied. 



290 



NICKEL STEEL FOR BRIDGES 



a 




^ 1 


E 


^^ 


H^^ 


vrtrv 


t.P 


a » 


« X 


Sco ^ 




B CB 








i-j ^ 


(T> r 




a> cs-^D 


d set 
the 
eflrs 
read 


o& 
1 n 


'ii 


CD ►I 


a (I 


rt-W 


B<6 


5'-" 


2S 






?3 


11 


(TO 


^'d 


gp 


^Sf 


?s 


<^ r^ 


Scrt; 






CU (t 


i? 


as- 


fi o 


•O ,^ 


•^ ^ 


2 tn 


0-3. 


p ■ 




® 2 


o 


B "^ 








N W 


&:■ 


<X> h^ 






cc 


> 


S5 


3 


§S 


« 


1^ a. 


» 




o 


> 
3 m 







3p 








oo 
















B*i-h 




CD O 




».>1 




Ptn 




^*k 




C-. 




B _. 




— CD 




p<<j 












CD 1 




^ CD 




S5 CD 








85 '1 








O 2 








i(>.^ 








55p 




is- 








9-8 




SS 




g*- 




p*o 












!?S 




«§• 






SP- 



88 



S^. 



W •- -> - -1 - -! -^ -1 -^ -I 
=■- ~. X — X — 3C CI (X 05 X 

S = o - 5 2 c o o o Q 
C C' = C O £ '^ o o o o 



tooi±; ocji 



oooooooooo 

to >-' p 'f O '►-' O '■-' cp >-' 



I M IS — ifi -J lE I-" • 



_o o p o 
»o o o o 



oooooooooo 



oooooooooo 
ooooooooc — 

OJOOOOOOOC 



g: 8 



CI OTtf^ »(i. t^O^k. M 



,*c ^jo to ,to 



~^^888: 



OOO OOOOOC. , 

OOO OOOOOOOOOOOC 



OOOOOOOOOOOO 
CO 03 io — o ►-' p '-' O i-' O 'l-' 

cooioi-^hfc*-i--.fe.m*i.k-*iEoo 



c» cji o; fONJi-' 

00 Oi 00 CJliOCD 



OOO OOOOOOOOOO- 
OO o o o o o O O O O Q o ■ 

H-l- OOQOOOOOOOC- 

0&OC003000000000* 



m u? ui w UT m »**. 



OOO o o 

OOO oo 



>b.i>s ts tc {s to 

cjiO'OOO'Xonccn.xciT'ac 
ooooopooooo 



: OOOOOOOOC 



o OOO oo- oooooooooo- 

CC CCCCM WM- to l-» O ">-' O '•-' O '-' O '►- "• 

00 oOTCO ►Ji-i- ooxvOx^'-jE x»^ao- 

00 ceok-" ow- ccoooc(»;Eoi I6x*'^-- 



oo 
too 



OOO oo- oooooooooo 



OOO oo 



> W to I-* ooo 
to CO tOCKCO 



QOOOOOOOOC 
OOOOQOOOOC 
OOOOOOOOC'C 



- »f^ (t- rfi. J^ CO CC CC *— 05 — CO N- 05 »-^ CC i-i 
105*. CCO-}*.OtOOtOOlOOtOOtO 



05 or 44- 



£l X 5' en SS 

*.. 00,{^ OS X 



O OOOOOOOOOOO 

to to i-i '•-'"•-• O '-' O -' O '■-' 

>— Q 00 OS *k CO en oo £H CO CII CO 

«o Si-' tolKcoocoocooco 



!» ^ O 

Op (D d erg 

r^^ "^ B£a 

(D ar* 



^2 



II ?* 






31 Hj 



5=^ 

■ > 



£,£ CD S o*? 



- 00 
t>tB 

s ** 



B o 
-•XI—.'O B P 

P c^ o ^o 



O O poo OOOOOOOOOOO- 

o ►- «»05 t-S coSo oooooS- 






t^ 



(X 



^ to 



GO 


• 


H 


t> 


H 


!?! 


B 


n 


r" 




1 


M 


1 


03 


f/> 




H 


td 


W 


K! 


M 




H 


K) 




5 


> 

O 


Kl 


hj W 


w 


> 


w 


w 


g 


c« 


> 




!?l 


G 


W 


►8 




§ 


no 


•-3 


H 


ffi 



K| 



O 



NICKEL STEEL FOR BRIDGES 



291 



!» toto 


to 


ftScteft 


•fc CO to 




«._co»o — 


S 3CX 


X 


' ' '5* 


tS tifO 


ts 


' ' 'P 



M 



03 

H 
O 



o:o:3cx0C3co!O!oiC5 



MtOtStstCltO'-'x'i-i^ 



H 
n 



iB^i^ r a 
"< o 

IS 19 tS tS to 10 i-* i_L .C 



:S *.fO X CiOicnx 

= oo o SQSS 
o oo o ocoo 



to «a5 05 Q0050 

o: oc 00 ecoiito 

as XX » ccisOTO 

O OQ O OOOO 

O oc O OOOO 



b 


en en 

■do: 


en 
<^ 

'cc 


cn_S*IeS 
b'-ib'co 






8g 

ocbo 


iO 


24.2 
20.3 
21.2 
25.8 




£io 

lob 


b 


12.5 
11.8 
10.7 
10.6 


CO *. ^B ^ o ;= 
OS f* en CO 013 


11.0 
9.5 

9.6 
9.5 


CO 


j^eo 
bo*. 


b 


lb. lb. CO lb. 

0501© 
'.-i*.bo05 


2) 
5' 




en 

•n 





cnenenenencncnencnos 



3000000000 
JOOOOOOOOO 



oo«o 000000 
coco.».oenxoiotoit^ 
»ocs»ocn->c:o-ixco 

8000000COQ 
oocooooco 



enenenencnencnencne:* 

co-jcncnc^'iXib.Xib.o 

o b b en Vo b en io b -~j 



i-'lOlOl-'iOlO'-'l-'— l-i 
OOI-'CCO'-COOXGD 

'lb. 00 ib. 00 -J 00 ie b "00 at 



ib.*.ib.*.iben4^CTib.en 

lb. _-* to -Q -I O X O -« CO 

cotOQoa)Ooocoib.oco 



> 




> 


-K 


> 


^ 


;^ 


r 


^ 


n 


3 






































ft 


K 


"T. 


Tq 


5 


jq 


IT 


n 


T 


K 




s 


» 


J.; 




2 


s: 


a 




3 


a> 


» 


rc 


s: 


fD 


P 


a> 


9: 


01 


p 








K 
^ 


p. 




fi. 




C, 





r» 



.to 



5'2. 




r^p- 


«- 




P5 


g-c: 


S'tJ 


CD Et 


Ho 




a§ 


PQ en 



o m5 

rt- "1 O' C. 



® Ts 2 2 

B (6 en B D. 


















s 




•fl 




> 




» 


M 


CO 


►«^ 





H 


!2! 


W 





!>► 


"9 


P3 






H 


^ 




> 


OQ 






ts 


f 


P3 


B 


P 


H 


1 


w 




WJ 


fei 


CO 












w 


!?l 


w 




r 


r/3 




►a 


f/J 


w 


H 





1 » 


K 


1 r 






^ 




«! 




> 




2 




D 




t^ 




C 




r 



CO 






292 



NICKEL STEEL FOR BRIDGES 



TABLE 67. — Tensile Tests on A. A. S. M. Specimens. Eye-Bar 

16 BY 2-ix. 

6 and 8-in. bars— Heat No. 17 749 ; 











Load, in Pounds per 






Original 






Square Inch. 


o 
■e-e 


Cut 
from: 






Fractured 
section, 
in inches. 






2g 


Section, 


Area, 
in 


Yield 


Ultimate 


Ed 




in 
inches. 


square 
inches. 




point. 


strength. 


6by 1-in. 


Diameter = 0.959 


0.722 


Diameter — 0.655 


62 300 


104 300 


59.7 




Diameter = 0.962 


0.727 


Diameter = 0.700 


56 400 


103 800 


54.9 


ifc 


1.255 by 0.985 
1 250 ' 99S 


1.336 


0.900 by 0.665 
0.930 '■ 0.695 
0.925 " 0.675 




102 400 




11 


1.344 
1.246 




98 600 
101300 




11 


1.265 ' 


0.9a5 


■ ■ ■ 62 m ' ' 


"'^\M 


11 


1.255 ' 


995 


1.249 


0.930 " 0.680 


55 200 


98 700 


55.9 


" 


1.250 • 


0.985 


1.231 


0.905 " 0.685 


56 900 


104 500 


54.5 


" 


1.245 ' 


0.990 


1.333 


0.925 " 0.700 


53 700 


99 800 


52.8 


8 by 2-ln. 


0.760 ' 


1.985 


1.509 


0.505 " 1.465 


*61 600 


98 700 


62.4 




0.780 ' 


1.990 


1.553 


0.515 '• 1.490 


51500 


94 500 


54.5 


11 


0.755 ' 


1.980 


1.495 


0.475 '• 1.480 


54 700 


93 800 


.58.3 




0.755 ' 


1.975 


1.491 


0.510 ' 


1.480 


51000 


92 600 


55.1 


11 


0.750 ' 


1.990 


1.493 


0..500 ' 


1.470 


+60 900 


99 900 


61.0 


u 


0.765 • 


1.985 


1.518 


0..530 ■■ 


1.510 


54 000 


96 800 


55.8 


11 


0.755 ' 


1.985 


1.499 


0.515 ' 


1.495 


58 700 


101 800 


57.7 


u 


0.755 ' 


1.990 


1.502 


0.555 ' 


1.550 


54 600 


98 900 


55 2 


11 


0.740 ' 


1.985 


1.469 


0.520 ' 


1.535 


59 900 


103 500 


57.9 


il 


0.760 ' 


1.985 


1.509 


0.510 ' 


1.510 


53 000 


96 100 


55.2 


u 


0.765 ' 


1.980 


1.515 


0.525 ' 


1.460 


50 400 


95 400 


.52.8 


11 


0.755 ' 


1.980 


1.495 


0.500 • 


1465 


50 700 


92 300 


54.9 


16 by 2-iD. 


0.760 ' 


2.030 


1.543 


0.525 ' 


1.570 


■ 59 600 


104 000 


57.3 




0.760 ' 


2.035 


1.539 


0.550 ' 


1.575 


55 900 


103 700 


53.9 


It 


0.755 ' 


2.025 


1.529 


0.515 ' 


1.550 


58 500 


103 100 


56.7 


^' 


0.760 " 2.025 


1.539 


0.540 " 1.550 


55 200 


102 700 


53.8 



Pieces cut out from 6 by 1-in. Bar, Bl, were turned to 1 in. in diameter and used for 
" " " " 6 by 1-in. Bars B2, 3 and 4, and piece TSLEBS2, were tested on 
" •' 8by2-in. and 16 by 2-in. bars, except TSLEBS2, were tested on 
* Autographic curves are abnormal at yield point. 
+ Broke IJ^ in- from end gauge mark. 



NICKEL STEEL FOR BRIDGES 



293 



Materul — Nickel Steel — ^Cut from 6 by 1-in., 8 by 2-in.^ and 

Eye-Bars. 

16-in. bars— Heat No. 17 673. 



Percentage of 
Elongation in: 


o 


t- 6 


Speed of Breaking. 


°d 






a 


<D t- 




a 01 






.2 cs 
2 £ 


ract 
ictu 




.2E 
S'3 














Remarks. 










5« 


e3 n 






S <D 




2 in. 


4 in. 


6 in. 


Sin. 


PS 




At yield point. 


At break. 


S^ 




37.0 


26.5 


21.7 


18.8 


53.3 


S i^ Cup. 


Very slow. 


1 in. in 2 min. 


All 


Original. 


35.0 


25.5 


21.0 


18.8 


47.0 


S Jl Cup. 


" " 


'• 


half 


Annealed. 


40.5 


26.5 


21.3 


19.0 


.51.6 


SCup. 


1 in. in 10 min. 


1 in. in 1 min. 


way 


Original. 


37.5 


25.5 


21.2 


18.5 


48.1 


S ?4 Cup. 


>' 




between 


Annealed. 


41.0 


28.3 


23.0 


20.6 


49.9 


SAng. 


" 




edge 


Original. 


38.5 


27.0 


22.3 


19.5 


49.4 


SCup. 


" 




and 


Annealed. 


40.5 


26.5 


21.7 


19.3 


49.6 


S Ang. 


" 




middle 


Original. 


40.0 


28.3 


22 7 


19.5 


47.4 


8^4 Cup. 


" 




of bar. 


Annealed. 


40.0 


30.0 


23 [3 


20 3 


51.0 


SAng. 


1 in. in 20 min. 


1 in. in 2 min. 


Middle. 


Original. 


40.0 


28.5 


23.3 


tso.o 


50.6 


S Cup. 


" 




" 


Annealed. 


35.0 


31.0 


25.7 


22.4 


53.0 


S Irrpg. 


'• 




Edge. 


Original. 


45.0 


81 .5 


25.3 


21.8 


49.4 


S Ang. 


" 




" 


Annealed. 


44.0 


31.2 


25.7 


22.4 


50.8 


SAng. 


" 




Middle. 


Original. 


42.0 


20.0 


23.3 


20.3 


47.3 


SAng. 


" 




" 


Annealed. 


42.0 


30 5 


25.0 


22.0 


48.6 


s Cup. 


• ' 




Edge. 


Original. 


37.0 


28.5 


23.7 


20.8 


43.7 


SCup. 


'■ 






Annealfd. 


39.0 


26.0 


20.0 


+17.5 


45.7 


SAng. 






Middle. 


Original. 


40. 


29.5 


24.0 


20.8 


49.0 


SAng. 






" 


Annealed. 


45.5 


32.0 


25.8 


22.0 


49.5 


SCup. 


1 in. in 10 min. 


1 in. in 1 min. 


Edge. 


Original. 


45.0 


31.0 


25.3 


22.0 


51.0 


S Ang. 


1 in. in 20 min. 


1 m. in 2 min. 


" 


Annealed. 


42.0 


29.5 


24.3 


20.5 


46.6 


S Cup. 


•i' 




Middle. 


Original. 


34.0 


24.5 


22.0 


19.3 


43.0 


SCup. 


'• 


•' 


" 


Annealed. 


42.0 


29.5 24.0 


21.0 


47.8 


SAng. 


'• 


" 


Edge. 


Original. 


37.0 


27.5 


22.3 


19.5 


45.6 


SAng. 








Annealed. 



Coeflacient of Elasticity determinat ion. 

150 000-lb. Kiehle machine at Rieble Bros. Co. 

20'JOOO-lb. Olsen machine at Drexel Institute. 



291 



NICKEL STEEL FOR BRIDGES 



?£! 



21 



>0>0>Ott»0>0>0>0>0 

re — • ^ —'CD ■-• C^ w '^ ■-. tT) — • (^ — ■ fT — ■ 

^3?:C83_=^B83Bp_Sg_Bj:B 

S'S.ffiS.x'^a £. (t" ^ 7 £. S" £. fT — 






IS tS 1-^ tS >-' tS IS iO H-" 1-'-^ 

IS 00 oc *. ^ ac' "ii *.^"ii "li"^"^ 

w O C: O C5 (X 



CO CO 'X ►- ffl "o ■— C O 05 X -* O to to 5D 



tS IS ►-»«>- JO t050-''-'i- 

CO i-i l-> — -.1 ^ I— ii H-i 
(X IS~^ >^ X OC 0^ i-^IS 



.CO •-' CO M CO 



?= B 9= „ g ?P5 

?> -<»s» S B ^te 
ra t^B ■ 3 



<K 



a>9 
.13 



B fS 

tf SO 
re rf- 



(H) Bfi 



B D 

-■a'B 

a aio<5 

ai? B 










bd 






w 






iz: 



















!z! 













H 






w 






w 


H 




H 


a> 






tc 







CO 




!2! 


h=l 




hj 






f 


» 




> 


CD 




!z: 





W 


r/3 


|-i 


K! 


'V 




M 


M 


^ 


td 


g 


Ou 


fO 


M 


<-i 




"y, 




g 


w 


0" 


> 




G 


W 


S 


H 


fD 


> 


•II 


» 


f 


Ki 


a- 


1 





(-*• 








(J<5 





05 


^ 


w 


ts 


P 




^ 





*■ 




!3- 


r/3 


r-i 


t-( 


H 


1— « 


ft) 


M 


'-«i 




M 


%I 


^ 


r 


on 


Oi 






a 




W 







►<^ 


M 
o- 




fcO 



hj 



^ 



$0 



NICKEL STEEL FOR BRIDGES 

APPENDIX B. (PART I 



295 



TABLE 59. — Chemical Analyses and Specimen Test Kesults of 
Material from Which Full-Sized Tests Met Specifications. 
B-1 200 — Blackwell's Island Bridge — Nickel-Steel Eye-Bars. 





Chemical Analysis 




Unannealed. 




Annealed. 


Heat 
No. 


C. 


P. 


s. 


Mn. 


Nl. 


i'g 




id 

§•2 


1 

o 

a 


11 


II 


i . 
§•2 


d 
o 

3 














W" 


5| 




Oh 


3- 


Dm 


S"^ 


•a 
02 


15 119 


0.36 


0.010 


0.027 


0.75 


3.36 


56 980 


100 500 


18.7 


35.9 


50 4.50 


94 080 


27.0 


50.5 


15 121 


0.32 


0.014 


0.029 


0.65 


3.36 


55 740 


95 260 


21.2 


37.1 


50 880 


87 740 


26.2 


50.9 


15 121 


0.32 


0.014 


0.029 


0.65 


3.36 


55 740 


95 260 


21.2 


37.1 


50 880 


87 740 


26.2 


50.9 


15 121 


0.32 


0.014 


0.029 


0.65 


3.36 


55 740 


95 260 


21.2 


37.1 


50 880 


87 740 


26.2 


50.9 


16 122 


0.37 


0.016 


0.031 


0.70- 


3.34 


57 500 


99 350 


19.5 


30.8 


55 100 


97 340 


19.7 


54.4 


16 122 


0.37 


0.016 


0.031 


0.70 


3.34 


57 500 


99 350 


19.5 


30.8 


55 100 


97:340 


19.7 


54.4 


16 068 


0.32 


0.015 


0.030 


0.65 


3.22 


57 230 


96 510 


20.0 


34.5 


51 840 


87 470 


24.2 


44.3 


1088 


0.38 


0.012 


0.029 


0.60 


3.36 


55 200 


100 4(H) 


20.0 


29.2 


53 360 


90 8.30 


21.2 


48.5 


16 245 


0.37 


0.010 


0.022 


0.74 


3.38 


60 650 


108 900 


17.5 


25.6 


54 020 


97 500 


20.0 


41.0 


16 230 


0.40 


0.013 


0.026 


0.67 


3.34 


57 890 


104 000 


16.2 


25.8 


53 720 


91 520 


25.0 


47.6 


16 257 


0..38 


0.012 


0.033 


0.72 


3.26 


59 .300 


104 5(X) 


16.2 


27.4 


53 160 


97 290 


23.7 


45.3 


16 313 


0.43 


0.014 


0.033 


0.75 


3.28 


56 180 


119 900 


15.0 


28.6 


51 260 


105 200 


21.0 


43.9 


2 139 


O.SH 


0.010 


0.028 


0.76 


3.28 


58 550 


102 200 


20.0 


32.0 


54 080 


96 200 


18.2 


46.1 


1 l(i5 


0.36 


0.010 


0.030 


0.73 


3.30 


58 940 


105 500 


19.2 


28.3 


51860 


86 400 


25.0 


44.9 


16 225 


0.41 


0.010 


0.034 


0.70 


3.36 


60 280 


110 100 


16.2 


34.0 


54 620 


101900 


20.5 


41.6 


1 114 


0.41 


0.012 


0.037 


0.62 


3.52 


59 270 


104 400 


14.0 


19.0 


52 400 


97 400 


21.2 


29.7 


16 227 


0.34 


0.016 


0.032 


0.70 


3.28 


55 380 


104 100 


20.0 


34.6 


55 640 


91 060 


21.2 


29.0 


16 227 


0.34 


0.016 


0.032 


0-70 


3.28 


55 380 


104 100 


20.0 


34.6 


55 640 


91 060 


21.2 


29.0 


16 222 


0.45 


0.013 


0.025 


0.70 


3.26 


64 780 


113 500 


18.1 


33.5 


.50 840 


101 200 


21.25 


46.4 


16 255 


0.41 


0.012 


0.026 


0.80 


3.30 


62 780 


115 400 


16.2 


31.8 


52 120 


104 200 


20.0 


34.2 


16 30H 


o.;m 


0.010 


0.030 


0.66 


3.28 


62 940 


107 200 


18.7 


25.4 


54 220 


93 360 


22.5 


4S.6 


16 248 


0.41 


0.012 


0.030 


0.80 


3.28 


66 360 


116 000 


15.0 


28.8 


60160 


104 800 


22.5 


42.8 


IV 260 


0.40 


0.010 


0.030 


0.78 


3.28 


55 240 


114 000 


16.0 


28.9 


53 780 


103 600 


18.5 


37.2 


16 316 


0.3H 


0.010 


0.028 


0.70 


3.40 


55 610 


95 140 


20.0 


38.7 


52 560 


97 940 


22.5 


44.1 


10 120 


0.41 


0.010 


0.035 


0.70 


3.40 


64 420 


119 800 


16.2 


23.7 


55 020 


103 200 


21.2 


42.2 


2 161 


0.38 


0.010 


0.035 


0.74 


3.26 


57 560 


106 200 


22.5 


45.4 


50 220 


100 900 


21.25 


33.6 


1 115 


0.43 


0.016 


0.025 


0.75 


3.32 


63 960 


110 100 


16.0 


28.9 


54 250 


104 200 


18.70 


38.2 


It; 2iK) 


0.39 


0.010 


0.035 


0.75 


3.36 


61260 


108.300 


17.2 


30.8 


54 010 


95 540 


23.7 


.51.3 


16 272 


0.40 


0.012 


0.025 


0.79 


3.25 


66 760 


111 600 


14.0 


20.1 


55 860 


102 600 


20.0 


40.2 


16 26:^ 


0.36 


0.012 


0.028 


0.74 


3.46 


65 250 


107 050 


17.5 


27.4 


48 040 


'M 070 


23.7 


45.0 


16 281 


0.44 


0.010 


0.026 


0.72 


3.28 


63 200 


116 .500 


15.0 


20.8 


55 61 K ) 


105 ,si)0 


22-5 


:32.2 


1() 280 


0.42 


0.011 


0.039 


0.69 


3.26 


67 640 


117 000 


13.7 


21.5 


51 460 


101 7(J0 


17.5 


31.5 


16 2!I2 


0.43 


0.010 


0.035 


0.80 


3.36 


66 080 


112 100 


14.5 


25.3 


55 140 


98 320 


21.2 


42.5 


16 239 


0.37 


0.010 


0.O30 


0.65 


3.46 


60 400 


101 600 


18.7 


.33.3 


53 950 


92 160 


21.2 


38.7 


16 23<.t 


0.37 


0.010 


0.030 


0.65 


3.46 


60 400 


101 600 


18.7 


33.3 


53 950 


92160 


21.2 


38.7 


16 3(13 


0.44 


0.010 


0.025 


0.80 


3.76 


60 620 


113 400 


17.2 


26.4 


55 850 


101500 


21.0 


43.9 


16 2:W 


0.37 


0.012 


0.033 


0.85 


3.30 


65 430 


110 200 


18.7 


32.1 


.50 440 


96 640 


15.0 


22.3 


10 0'.t6 


0.40 


0.010 


0.028 


0.62 


3.36 


55 760 


94 000 


20.0 


34.4 


49 670 


88:380 


21.2 


40.1 


16 231 


0.39 


0.012 


0.031 


SO 


3.30 


63 9(X) 


1 10 100 


16.0 


28.9 


55 400 


104 :3a) 


20.0 


30.0 


2 145 


0.36 


0.015 


0.030 


0.75 


3.28 


56 140 


105 400 


17.0 


35.2 


55 360 


97 700 


24.2 


50.5 


16 226 


0.37 


0.012 


0.0:30 


0.78 


3.26 


57 840 


109 la, 


18.7 


32.4 


55 630 


96 740 


23.7 


47.2 


16 2S5 


0.45 


0.011 


0.030 


0.78 


3.26 


67 12(1 


112 401) 


15.0 


27.4 


52 OSO 


ll)34(K) 


20.0 


37.4 


16 236 


0..36 


0.012 


0.035 


0.65 


328 


61 S(i(i 


ml (mil 


16.2 


25.0 


52 20; 1 


91 720 


21.2 


45.2 


16 322 


0.42 


0.010 


0.028 


0.77 


3.30 


66 (ISO 


llCi -.'INI 


16.7 


31.0 


5S Olio 


102 100 


20.0 


30.9 


Hi 223 


0.45 


0.011 


0.030 


0.75 


3.2S 


(iS 3(1(1 


lis 300 


15.0 


24.2 


54 250 


110 800 


21.0 


38.8 


16 224 


0.38 


0.011 


0.o:« 


0.72 


3.. 36 


(13 41(1 


11)5 400 


15.0 


17.3 


52 130 


95 620 


18.25 


20.3 


10 115 


0.40 


0.012 


0.032 


0.H7 


3.2(i 


64 4S(I 


105 200 


14.5 


18.8 


55 980 


105 900 


20.0 


35.3 


1 110 


o.as 


0.010 


O.OIil 


0.72 


3.46 


I'jii mill 


1 1 )2 2( )i 1 


18.7 


28.2 


51540 


94 800 


21.2 


41.0 


1 123 


0.43 


0.916 


0.030 


0.72 


3.32 


62 2i;ii 


110 3iHi 


17.2 


30.9 


62 140 


102100 


21.2 


41.0 


16 258 


0.37 


0.010 


0.026 


0.73 


3.28 


."i7 4-.'0 


101 HOD 


18.7 


32.9 


53 000 


94 200 


23.7 


40.6 


10 111 


0.10 


0.010 


0.030 


0.64 


3.:i4 


>'i2 3(i(i 


110 200 


17.06 


32.2 


56 050 


98100 


21.2 


87.6 


2 165 


0.37 


0.013 


0.037 


0.80 


3.2N 


Ii2 CiOd 


113 200 


17.5 


15.9 


51 600 


99 480 


20.0 


46.9 


10 116 


0.37 


0.014 


0.029 


0-70 


3.36 


61 ;",7il 


l(iii:iO(; 


20.0 


31.8 


54 270 


84 100 


25.0 


49.1 


16 311 


0.37 


0.010 


o.o:« 


0.75 


3.36 


(;2 4:.ii 


1 1 1.- ^1 II 1 


18.7 


■S4A 


54 880 


97 620 


21 .2 


45.5 


Iti 256 


0.40 


0.012 


0.022 


0.73 


3.28 


67T'"> 


1 1 ■.' ■-'■ 1' 1 


14.5 


24.1 


53 250 


101 600 


22.5 


47.3 


16 851 


0.38 


0.010 


0.028 


0.64 


3.48 


56 2:^,11 


11 HI ,-,1111 


20.0 


32.7 


53 980 


a5 400 


26.2 


53.0 


16 861 


0.39 


0.010 


0.030 


0.69 


3.34 


55 9U0 


103 800 


18.7 


30.7 


52 900 


94 440 


24.7 


48.5 


14 566 


0.41 


0.011 


0.031 


0.65 


3.60 


60 900 


105 900 


19.7 


38.3 


56 560 


97 020 


23.5 


40.1 



2i)G 



NICKEL STEEL FOR BRIDGES 











TABLE 


59.— 


(Oon 


tinned.) 












Chemical Analysis. 


Unannealed. 


Annealed. 


Heat 














D -." 




g 




<E r." 




d 


No. 


C. 


P. 


s. 


Mn. 


Ni. 


o . 

3- 


~ a 
-.3 (c 

51 


§•2 


o 


3- 


51 


e3 . 

Ma 

a 




t 
1 

05 


13 571 


0.44 


0.012 


0.029 


0.78 


3.30 


65 950 


113 100 


16.2 


30.8 


61 680 


105 200 


21.3 


42.4 


14 543 


0.37 


0.013 


0.025 


0.80 


3.30 


64 280 


107 500 


18.7 


36.1 


52 320 


102 900 


20.0 


33.5 


16 859 


0.38 


0.010 


0.035 


0.66 


3.34 


55 500 


99 180 


20.0 


35.0 


58 340 


90 220 


24.7 


43.5 


14 559 


0.47 


0.013 


0.023 


0.72 


3.40 


59 350 


115 7O0 


13.2 


28.1 


48 220 


108 000 


16.3 


28.9 


16 268 


0.42 


0.010 


0.022 


0.78 


3.36 


55 620 


1 17 000 


14.5 


13.8 


56 200 


105 400 


18.7 


.39.1 


16 270 


0.42 


0.010 


0.036 


0.75 


3.36 


62 4(» 


115 2(J0 


16.2 


30.0 


51 000 


101000 


21.5 


45.6 


14 545 


0.40 


0.010 


0.036 


0.66 


3.26 


55 680 


104 100 


21.2 


36.9 


50 800 


95 060 


22.5 


45.1 


16 305 


0.44 


0.010 


0.025 


0.73 


3.38 


60 740 


11(1700 


17.5 


29.5 


52 040 


103 300 


22.0 


41.3 


14 555 


0.46 


0.010 


0.036 


0.65 


3.36 


57 300 


116 4(Kl 


15.0 


22.4 


62 620 


108 200 


20.5 


38.2 


14 530 


0.40 


0.010 


0.029 


0.63 


3.30 


55 220 


101 400 


23.2 


43.7 


50 100 


91 .560 


23.2 


54.5 


14 558 


0.39 


0.022 


0.031 


0.85 


3.40 


65 750 


110 4(.H) 


17.0 


.31.1 


58 820 


101 400 


22.2 


50.3 


12 738 


0.41 


0.010 


0.034 


0.72 


3.54 


62 020 


102 900 


18.7 


32.8 


52 960 


95 300 


22.5 


50.5 


14 552 


0.44 


0.016 


0.031 


0.63 


3.42 


63 120 


113 50ft 


12.5 


18.6 


56 720 


90 500 


33.7 


48.7 


1 116 


0.45 


0.015 


0.028 


0.80 


3.40 


66 320 


117 500 


13.7 


18.8 


59 900 


109 700 


21.2 


.34.2 


16 279 


0.44 


0.010 


0.031 


0.82 


3.26 


55 440 


112 500 


16.2 


29.6 


51840 


102 700 


20.5 


44.9 


12 041 


0.39 


0.012 


0.030 


0.75 


3.62 


61 160 


107O."i0 


18.2 


29.5 


56 400 


94 900 


25.0 


44.5 


12 059 


0.43 


0.013 


0.028 


0.70 


3.64 


59 940 


112 500 


12.5 


23.0 


55 380 


101 400 


18.7 


25.6 


12 054 


0.38 


0.017 


0.029 


0.63 


3.50 


65 630 


101 (iOO 


16.2 


23.1 


50 140 


86 700 


22.0 


39.8 


12 081 


0.40 


0.012 


0.029 


0.69 


3.40 


60 800 


107 400 


18.7 


26.7 


56 730 


96 320 


23.7 


41.6 


12 064 


0.43 


0.016 


0.032 


0.78 


3.58 


61 7(» 


102 300 


19.5 


34.1 


58 060 


96 800 


21.2 


43.7 


16 795 


0.40 


0.010 


0.0.30 


0.65 


3.58 


62 220 


103 000 


18.7 


26.3 


56 260 


95 200 


23.7 


46.7 


12 055 


0.38 


0.010 


0.03; 


0.69 


3.56 


62 320 


1(12 800 


18.5 


28.5 


56 360 


88140 


25.0 


49.5 


12 049 


0.38 


0.010 


(l.(«0 


0.78 


3.59 


64 170 


103 800 


17.0 


27.7 


57 160 


89 950 


25.0 


52.7 


16 306 


0.42 


0.01(1 


0.(130 


0.78 


3.32 


60 480 


109 800 


17.5 


33.6 


50 100 


98 960 


21.2 


44.3 


16 320 


0.42 


O.OIO 


0.027 


0.75 


3.32 


62 600 


110 TOO 


16.2 


32.0 


53 720 


95 740 


21.2 


42.8 


12 037 


0.39 


0.012 


0.03(1 


0.69 


3.46 


61 200 


1(1."] 000 


19.7 


28.5 


61 510 


93 010 


28.0 


,50.8 


16 262 


0.43 


0.011) 


022 


0.73 


3.39 


61 3(10 


116,500 


15.0 


31.3 


54 450 


104 200 


21.2 


40.4 


16 828 


0.36 


0.010 


o.o;ji 


0.76 


3.32 


64 490 


108 850 


17.0 


19.4 


58 220 


95 460 


21.7 


38.9 


16 330 


0.38 


0.014 


0.038 


0.70 


3.32 


60 950 


106 900 


18.7 


31.4 


57 200 


96 190 


21.2 


44.1 


14 548 


0.41 


0.010 


0.03(i 


0.73 


3.40 


61 800 


108 650 


20.0 


38.7 


59 180 


102 350 


22.5 


46.2 


16 791 


0.39 


0.010 


O.03O 


0.tj9 


3.52 


62 800 


101 ()0(l 


20.0 


32.2 


55 400 


93 300 


25.0 


48.7 


12 014 


0.39 


0.1112 


0.0.32 


0.66 


3.42 


55 760 


102 '.ICO 


20.0 


28.5 


53 200 


92 500 


21.3 


46.3 


12 080 


0.40 


0.010 


025 


0.75 


3.50 


59 000 


107 400 


20 


22.5 


55 190 


96 970 


34.5 


40.9 


14 743 


0.42 


0.010 


0.035 


0.71 


3.54 


59 930 


111 soo 


16.3 


27.4 


56 380 


100 600 


23.3 


47.1 


14 742 


0.42 


0.013 


0.032 


0.65 


3.46 


56 500 


104 900 


18.7 


29.3 


54 080 


86 240 


35.7 


45.1 


16 309 


0.37 


0.010 


0.032 


0.79 


3.26 


.57 160 


lo; 700 


20 


38.5 


51 260 


97 820 


26.2 


49.8 


14 747 


0.38 


0.012 


0.030 


0.69 


3.. 56 


60 140 


100 400 


20.7 


37.1 


,57 Odd 92 560 


27.5 


49.5 


12 737 


0.41 


0.012 


0.034 


0.65 


3.26 


57 600 


104 5(10 


18.7 


31.4 


56 940 93 480 


22.2 


47.0 


12 776 


0.40 


0.010 


0.029 


0.66 


3.62 


59 880 


102 900 


18.7 


29.6 


52 980l 94 460 


22.5 


45.1 


12 707 


0.41 


0.010 


0.029 


0.68 


3.30 


62 640 


106 70(1 


15.0 


31.2 


58 740 97 200 


23.7 


49.8 


1444 


0.38 


0.012 


0.034 


0.63 


3.30 


55 340 


98 700 


20.0 


29.7 


,52 28(1 >^(i 720 


23.7 


49.6 


12 261 


0.42 


0.010 


0.027 


0.70 


3.50 


61 900 


103 500 


21.2 


33.5 


52 940 9(i 500 


25.0 


42.9 


12 013 


0.41 


012 


0.032 


0.68 


3.40 


58 860 


103 500 


17.5 


27.3 


48 250 <ss240 


25.0 


45.3 


12 003 


0.37 


0.010 


0.034 


0.66 


3.64 


56 760 


102 700 


20.0 


28.0 


53 960 a") 720 


23.7 


40.7 


12 884 


0.45 


0.010 


0.037 


0.65 


3.60 


55 580 


109 3(10 


16.2 


26.2 


59 060 10(1. 500 


24.7 


41.6 


12 193 


0.39 


012 


029 


0.70 


3.48 


60 60(1 


10(1 100 


22.0 


38.3 


51 810| V2 010 


23.7 


45.6 


12 740 


0.38 


o.ois 


0.032 


0.69 


3.34 


61 020 


105 4(KI 


17.5 


34.0 


52 8(K) 93 460 


22.5 


50.7 


13 062 


0.37 


0.011 


0.032 


0.75 


3.46 


59 720 


101 ;io(i 


20.5 


40.9 


52ti4(l, 89 160 


23.7 


47.7 


12 340 


0.40 


0.0] •-' 


0.027 


0.62 


3.36 


62 840 


109 10(1 


15.0 


25.8 


53 940: 94 120 


33.2 


42.0 


12 709 


0.41 


0.010 


0.029 


0.68 


3.36 


(;i (100 


103 706 


18.7 


29.8 


48 330 96 540 


35.0 


46.1 


12 340 


0.40 


0.012 


0.037 


0.62 


3.36 


62 840 


109 40(1 


15.0 


25.8 


53 940j 94 120 


23.2 


42.0 


11434 


0.40 


0.012 


0.037 


0.65 


3.64 


58 340 


100 300 


19.5 


31.3 


53 700^ 89 920 


24.2 


41.9 


12 359 


0.39 


0.010 


0.032 


0.74 


3.50 


64 920 


101 700 


20.5 


33.7 


55 600 1 99 500 


23.7 


39.7 


12 358 


0.36 


O.OlO 


0.030 


o.t;o 


3.50 


.59 420 


97 160 


21.2 


34.1 


,57 4()0| 91 (i(.)0 


22.0 


38.1 


12 346 


0.40 


0.010 


0.027 


0.80 


3.62 


63 860 


104 500 


18.0 


34.0 


.59 170 94 060 


27.0 


49.1 


12 343 


0.37 


0.010 


0.027 


0.69 


3.64 


64 420 


108 800 


17.5 


34.3 


54 060 95 100 


2:12 


42.8 


12 452 


0.38 


0.010 


0.028 


0.76 


3.44 


62 9.50 


kk; 7(ki 


18.7 


25.1 


57 800 97 (;,hO 


23.2 


45.4 


12 727 


0.37 


0.010 


0.032 


.'•,8 


3.44 


57 32(1 


95 340 


20 


35.5 


48 030 88 820 


23.7 


48.3 


12 725 


0.39 


0.010 


0.025 


0.68 


3.61 


60 140 


1(12(K)(I 


17.5 


37.5 


57 500 96 KH) 


23.7 


42.8 


12 773 


0.40 


0.010 


0.030 


0.60 


3.34 


.58 70(! 


102 6(KI 


18.7 


3:i.o 


55 42(1 90 700 


26.0 


50.0 


12 728 


0.36 


0.011 


0.0 v'T 


0.60 


3.36 


59 860 


103 300 


21.2 


35.1 


57 480 93 320 


25.0 


49.1 


12 539 


0.39 


0.011 


031 


0.77 


3.40 


,59 940 


100 800 


20.0 


25.7 


51 5t)0 91 960 


25.5 


44.1 


13 450 


0.36 


0.01(1 


o.o:!o 


0.54 


3.36 


.59 940 


99 450 


16.2 


30.1 


54 410 91 800 


23.7 


43.7 


11437 


0.37 


0.010 


0.029 


0.59 


3..-)3 


59 600 


101 9(10 


18. 2 


26.4 


,59 9." 95 920 


21.7 


34.8 


13 429 


0.36 


0.010 


0.033 


0.63 


3.. 50 


61 040 


99 360 


18.7 


30.4 


,55 37(1 93 080 


24.5 


45.9 


14 553 


0.43 


O.OIS 


0.032 


0.73 


3.34 


63 740 


112.8(H) 


14.7 


26.8 


57 72(1 103 SIX) 


18.7 


28.9 


18 345 


0.43 


0.012 


0.033 


0.69 


3.50 


61340 


108 400 


16.7 


30.1 


56 106| 98 760 


22.0 


39.2 



NICKEL STEEL FOR BRIDGES 



297 



APP»K]SrDlX K. (PAKT II.) 

TABLE 60. — Eesult of Tests That Meet Specifications. Black- 
well's Island Bridge — Nickel-Steel Eye-Bars. 



6 


Section. 




— "S « 
" != !3 


Ultimate tensile 

strength, in pounds 

per square inch. 


a o 

Sm 

a., 

.2 

■33 


i. 

CS 

So 

D 

'■3 " 


Character of fracture. 


15 119 


16 X 2 


3.36 


54 260 


87 710 


6.80 


12.3 


Crystalline. Broke in head. 


15 121 


16 X 1 7/8 


3.36 


52 460 


80 270 


10.00 


36.1 


Silkv. 


15 121 


16 X 1 15/16 


3.36 


50 590 


85 841 


11.10 


40.5 


950/6 silky. 


15 121 


16 X 1 7/8 


3.36 


51 7K0 


88 280 


10.32 


28.6 


Cup. 05"/, .silky. 


16 122 


16 X 1 7/8 


3.H4 


51 320 


8(i 300 


6.55 


8.7 Crvstalliiie. Broke in head. 


16 122 


10 X 1 7/8 


3.34 


52 740 


90 080 


7.00 


11.8 frvstalliiie. Brolve in head. 


16 068 


16 X 1 7/8 


3.22 


51 1.50 


89 840 


10.60 


10.3 'm'\, e'-ystalline and 40% silky. 
















Broke in head. 


1 088 


16 X 1 7/8 


3.36 


49 a50 


82 540 


14.10 


33.8 


Silkv: irregular. 


16 245 


16 X 1 7/8 


3.36 


49 810 


87 970 


14.77 


39.6 


" *^"P- u , 


16 230 


16 X 2 


3.34 


51 280 


90 120 


9.22 




Fine granular. Broke in head. 


16 257 


16 X 2 


3.26 


50 660 


89 470 


13.21 


44".5 


Silky cup. 


16 313 


16 X 2 1/8 


3.28 


50 072 


98 980 


10.55 


38.9 


Irreg. silky. Trace crystalhne. 


2 139 


16 X 2 1/16 


3.28 


52 990 


88 310 


18.44 


38.8 


Silky. 90",, cup. 


1 105 


16 X 1 7/8 


3.64 


48 051 


85 535 


13.05 


35.3 


'■ angular. 


16 225 


16 X 21/16 


3.36 


48 102 


88 137 


13.50 


21.4 


Irreg. silky and fine granular. 


1 114 


16 X 2 


3.52 


48 296 


85 446 


16.61 


26.9 


Silky W% cup. 


16 227 


16 X 2 


3.29 


51 314 


86 357 


14.88 


4.3 


" 70% " 


16 227 


16 X 2 


3.28 


49 340 


88 623 


11.55 


38.9 


■' angular. 


IS 222 


16 X 2 


3.20 


51 020 


95 392 


12.11 


40.4 


•' " 


10 255 


16 X 1 3/4 


3.30 


48 170 


98 840 


10.00 


.33.9 


Fine granular. 


10 308 


16 X 1 15/16 


3.28 


49 315 


82 410 


17.61 


44.7 


Silky 14 cup. 


16 248 


16 X 1 7/8 


3.28 


49 1.50 


91 100 


16.16 


29.1 


" angular. 


16 2ai 


16 X 1 7/8 


3.28 


49 210 


95 380 


12.27 


37.2 


" '• 


16 316 


16 X 1 15/16 


3.40 


4h 150 


82 950 


12.11 


39.7 


" irregular. 


10 120 


16 X 1 7/8 


3.40 


48 400 


84 500 


12.83 


35.9 




2 161 


16 X 1 3/4 


3.26 


51 240 


8:-5 400 


10.83 


48.8 


" cup. 


1 115 


16 X 2 1/16 


3.32 


49 100 


98 3.50 


11.22 


.36.7 


'• square. Trace granular. 


16 ^90 


16 X 1 7/8 


3.36 


49 300 


92 080 


16.00 


44.0 


'■ cup. 


16 272 


16 X 2 1/8 


3.28 


51 U70 


98 000 


18.38 


8.7 


Broke in head. 


16 263 


16 X 2 


3.46 


53 000 


87 600 


15.94 


44.2 


Silkv 1^ cup. 


16 281 


16 X 2 


3.28 


49 290 


94 890 


10.55 


28.3 


Irreg. 800b crystalline, bal. 
silky. 


16 280 


16 X 2 


3.26 


49 040 


96 700 


12.16 


27.4 


Square. Near silky. 


16 292 


16 X 2 


3.36 


52 300 


95 880 


13.90 


38.4 Silkv cup. 


16 239 


16 X 1 7/8 


3.26 


48 130 


84 160 


12.66 


21.8 7.5",, crvstalliiie, bal.irreg. silky. 


16 239 


16 X 1 7/8 


3.26 


48 150 


89 020 


14.00 


.36.5 


Silky irregular. 


16 :i03 


16 X 1 7/8 


3.26 


48 110 


80 000 


9.66 


28.7 


759o crystalline, bal. silky. 


16 238 


16 X 1 3/4 


3.30 


50 m) 


91 000 


15.94 


41.3 


Silky cup. 


10 096 


16 X 2 


3.36 


49 220 


90 020 


11.77 


34.2 


" irregular. 


16 231 


16 X 2 1/8 


3.30 


48 -230 


95 000 


I0.8:i 


40.5 


" 


2 145 


16 X 2 1/16 


3.28 


53 110 


90 670 


13.77 


6.8 


" 25%) angular, bal. square. 


16 226 


16X2 


3.26 


48 920 


93 320 


11.11 




Too long to pull to fracture. 


16 285 


16 X 2 1/16 


3.26 


39 960 


90 480 


13.00 


4415 


Irreg. 00% silky, bal. crystal- 
line. 
Silky angular. 


16 236 


16 X 1 7/8 


3.28 


49 3.30 


85 250 


14.80 


44.5 


16 322 


16 X 1 15/16 


3.30 


48 270 


93 .500 


9.06 


41.4 


" " irregular. 


16 22:5 


16 X 2 1/16 


3.28 


48 250 


91 010 


13.33 


41.9 


Irregular silky. 


16 224 


16 X 2 


3.36 


49 85(1 


89 200 


12.16 


Too long to pull to fracture. 


10 115 


16x2 


3.26 


49 200 


95 760 


13.16 


32.7 Irregular silky. 


1 110 


16X2 


3.46 


49 090 


90 390 


17.11 


»1.2 Irreg. 5C% silky, bal. granular. 


1 123 


16 X 1 7/8 


3.32 


50 530 


91 810 


11.05 


17.0 Square granular. 


16 258 


16 X 1 7, 8 


3.28 


48 260 


83 630 


11.50 


31.8 Irregular silky. 


10 111 


16 X 1 7/8 


3.34 


51 070 


85 490 


9.44 


Granular. Broke in head. 


2 165 


16 X 1 7/8 


3.28 


50 140 


83 300 


14.27 


43.0 


Silky irregular. 


10 116 


16 X 1 7/8 


3.. 36 


48 120 


80 480 


10.94 


26.6 


angular. 


IC 311 


16 X 1 7/8 


3.36 


50 260 


85 980 


9.77 


38.8 





298 



NICKEL STEEL FOR BRIDGES 



TABLE m.— (Continued.) 



Section. 



16 256 
16 851 
16 861 
14 566 

13 571 

14 543 
16 859 
14 559 
16 268 

16 270 
14 545 

16 305 
14 555 
14 530 

14 558 
12 738 
14 552 
1 116 
16 279 
12 041 
12 059 
12 054 
12 081 

12 064 
16 795 
12 055 
12 049 
16 306 
16 320 
12 037 
16 262 
16 328 
16 330 
14 548 
16 791 
12 014 
12 080 
14 743 
14 742 
16 309 
14 747 
12 737 
12 776 
12 707 
1 444 

12 261 
12 013 
12 003 
12 884 
12 193 
12 740 
12 062 
12 840 
12 70^ 
12 340 
11 434 



6X2 

6 X 1 778 

6X17/8 

6X2 1/16 

6x17/8 

6X2 

6 X 1 15/16 

6X2 

6X2 

6 X 1 15/16 
6 X 1 13/16 

6X2 
6x17/8 
6 X 1 15/16 

6X17/8 
4 X 1 11/16 
4 X 1 11/16 
6X2 1/16 
6X2 1/16 
6 X 1 13/16 
6 X 1 13/16 
6 X 1 15/16 
6 X 1 13/16 

6 X 1 13/16 

6x2 1/8 

6 X 1 15/16 

6X13/4 

6x21/8 

6x21/8 

6x2 1/8 

6X2 1/16 

6x2 1/8 

6X21/8 

4X17/8 

6x21/8 

6X2 

6X2 

6x17/8 

6X2 1/16 

6x2 1/8 

6X13/4 

6X2 

6X2 

6X2 1/16 

6X2 

6x2 1/8 
6x2 1/8 
6x2 1/8 
6x2 1/8 
6 X 1 15/16 
6X2 

6 X 1 15/16 
4 X 1 15/16 
6X2 
6X2 1/8 
6 X 1 15/16 







w 








•s 


a 




tx— < 


3 ftd 

.5 Ki'" 


gft-S 


^^M 


^^'d (U 


*■" s 


g-3 


1^1 


4J a t- ■ 


eu 




0EK 

98 060 


8.28 


54 040 


3.48 


48 290 


83 520 


3.34 


48 060 


86 180 


3.60 


57 050 


87 770 


8.80 


48 520 


90 030 


8.30 


48 160 


87 290 


3.34 


48 890 


82 740 


3.40 


48 120 


90 780 


3.86 


52 970 


87 520 


3.86 


48 050 


90 920 


8.26 


48 110 


84 970 


3.38 


53 200 


88 030 


8.36 


52 870 


95 010 i 


8.30 


51 340 


86 050 


8.40 


48 200 


91 920 


3.54 


50 400 


85 470 


3.42 


48 270 


91 760 ; 


3.40 


48 240 


93 900 


3.36 


56 380 


94 480 ; 


8.62 


58 000 


89 860 


3.64 


48 980 


88 870 


3.50 


50 080 


86 170 


3.40 


51 320 


87 980 


3.58 


52 160 


92 150 


3.58 


48 290 


86 180 


3.56 


48 140 


87 400 


8.54 


52 210 


87 620 


3.32 


48 230 


82 650 


3.36 


48 310 


84 470 


3.46 


53 490 


89 560 


3.34 


48 240 


87 680 


8.33 


48 210 


82 730 


3.82 


48 170 


81 670 


3.40 


53 880 


91 890 


3.52 


49 230 


82 250 


3.42 


48 040 


85 580 


3.50 


48 010 


87 280 


8.54 


55 900 


98 000 


3.46 


51 350 


90 180 


8.26 


55 930 


92 800 


3.56 


54 112 


87 810 


3.46 


55 690 


86 320 


8.62 


54 820 


90 220 


3.30 


56 270 


89 130 


3.30 


52 860 


83 020 


3.50 


50 000 


85 500 


3.40 


52 050 


86 170 


3.64 


54 390 


88 970 


3.44 


49 160 


80 360 


3.48 


4K 240 


81 330 


3.34 


48 460 


87 190 


3.54 


60 570 


87 790 


3.62 


48 260 


87 5t)0 


3.36 


51 130 


87 670 


3.86 


52 080 


89 380 


3.64 


51 170 


87 420 



■s a 
a o 



11.94 
12.16 
9.11 
8.16 
10.55 
10.94 
11.22 
9.77 
6.05 

10.33 
9.16 

6.00 
7.55 
7.38 

13.66 
13.05 
10.83 

9.72 
12.05 
14.27 

9.44 
10.55 

7.22 

10.00 

10.94 

10.88 

11.60 

8.44 

9.88 

6.44 

13.88 

9.11 

12.44 

8.83 

9.16 

10.27 

10.11 

10.77 

9.88 

7.28 

11.05 

12. ,33 

7.22 

6.50 

11.27 

12.55 

6.88 
10.38 
10.00 
11.16 
12.77 
11.00 
12.05 
8.50 
7.66 
6.83 



O «8 



.2^ 



Character of fracture. 



31.9 
34.2 
44.5 

43 '.4 
45.4 
44.5 
37.2 



44.4 
32.2 



38.8 
43.8 
37.9 
32.8 
38.5 
32.7 
34.9 
41.6 



33.3 
41.9 
41.1 
43.9 
27.4 
35.5 

43!5 

88.2 

45.8 

46 .'3 
38.9 
39.8 
34.2 



42.0 
33.9 



28.2 

42.9 

46!2 
37.2 
34.4 
43.0 

45;3 



Silky irregular. 



Granular. Broke in head. 
Irregular. 40% silky, hal. grau. 
Irregular. 60% silky, bal. gran . 
Silky angular. 
20% silky, bal. crystalline. 
50% silky, bal. crystalline. 

]3roke in head. 
Silky irregular. 

Irregular. 30% silky, bal. crys- 
talline. 
Granular. Broke in head. 
Granular. Broke in head. 
30% silky, bal. crystalline. 

Broke in head. 
Silky angular. 
Silky 14 cup. 
Silky irregular. 
20% silky, bal. cryst. 
Silky irregular. 

" angular. 

" irregular. 
1 70% silky, bal. cryst. 
300/0 Broke in 

head. 
Silky irregular. 

" cup. 

" irregular. 

" J^cup. 

" square. 

" irregular. 
Granular. Broke in hea<l. 
,80% silky, bal. cryst. 
Silky irregular. 

" 14 cup. 
Crystalline. Broke in head. 
Silky irregular. 

50% silky and bal. cryst. 
|10% " - 

Fine crystalline. Broke in head. 
Crystalline. Broke in head. 
Silky irregular. 

" angular. 
Granular. Broke in head. 

Crystalline. 

Angular. 14 silky ami bal. 

crystalline. 
Silky irregular. 
Granular. Broke in head. 
Silky irregular. 

" ' angular. 

" irregular. 

" angular. 
Granular, Broke in head. 
Silky aiit,'iilar. 

8OO0 crystalline. Broke in head. 
Granular. Broke in liead. 



NICKEL STEEL FOR BRIDGES 

TABLE QO.— (Continued.) 



299 







•s 


a 


M S o 

fl o a 


©00 


OS d5 




d 




bD-i 
03 9i 


.5 01 




S a 
a o 


O (S 




i 


Section. 




— 12 » 
o - 


09 


K 


Cliiiiacter of fracture. 


12 359 


16 X 1 15/16 


3.50 


57 260 


96 930 


6.00 




Granular. Broke in head. 


12 358 


14 X 1 15/16 


3.50 


48 140 


83 550 


11.66 


35 '.5 


[rregular. 60% cryst. and bal. 

silky. 
Silky 'j^ cup. 


12 346 


14 X 1 7/8 


3.62 


58 210 


97 920 


12.16 


42.7 


12 343 


16 X 1 3/4 


3.64 


57 500 


94 260 


6.00 




Granular. Broke in head. 


12 452 


14 X 1 3/4 


3.44 


56 000 


95 090 


9.06 


32] 6 


Silky irregular. 


12 727 


16 X 2 1/16 


3.44 


53 440 


85 420 


6.55 




75% silky, bal. gran. Broke in 
head. 


12 725 


16 X 2 1/16 


3.64 


50 320 


85 910 


9.78 


44.7 


Silky irregidar. 


12 773 


16 X 2 1/16 


3.34 


49 030 


83 840 


10.94 


40.4 


" cup. 


12 728 


16 X 21/16 


3.86 


56 680 


85 800 


14.28 




" irregular. Broke in head. 


12 539 


14 X 1 15/16 


3.40 


52 270 


89 620 


9.00 


37! 8 


" cup. 


12 450 


14 X 1 15/16 


3.36 


52 920 


m (180 


9.11 


4.1 


" angular. 


11 437 


16X2 1/8 


3.53 


53 030 


89 070 


6.05 




Granular. Broke in head. 


12 429 


16 X2 


3.50 


50 060 


89 740 


12.83 


33!4 


Silky square. 


14 553 


14 X 1 7/8 


3.34 


53 090 


87 560 


6.55 




Crystalline. Broke in head. 


12 345 


16 X 2 


3.50 


51.020 


86 420 


14.77 


4i;6 


50% silky and 50% fine crystal- 
line. 














1 



300 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

DISCUSSION. 



Mr. Fowler. Charles Evan Fowler, M. Am. Soc. C. E. (by letter). — The engi- 
neering profession and the metallurgists engaged in the manufacture 
of structural steel are certainly to be congratulated upon the extensive 
and careful investigations made by Mr. Waddell into the use of 
nickel steel for structural purposes, and also upon the work that has 
been done by F. C. Osborn, M. Am. Soc. C. E., in assisting in this 
valuable work. 

The writer was one of the first to adopt the use of soft medium 
steel for structural purposes, as covered by "General Specifications for 
Steel Roofs and Buildings," and is still of the belief that, for a great 
many years, nothing better will be found for short-span bridges. 

The small saving in cost by using either medium steel or nickel 
steel, would hardly be a valid reason for making a change, and, even 
while it might be considered advisable by some engineers to use nickel 
steel for medium-length spans, it will undoubtedly be many years before 
soft medium and medium steel are entirely displaced in the building 
of short and medium-length spans. 

The great ease with which members can be fabricated from soft 
steel, and the reliability of the structures with only the usual amount 
of fair reaming in the shop, are almost unanswerable arguments in 
favor of continuing the use of this metal for short spans. 

It is to be presumed from the data in the paper that the difference 
in cost of fabrication, as between medium steel fully reamed, and 
nickel steel, is very small; so that, for spans of considerable length, 
and for very long spans, there can be little question that if nickel 
steel proves to be as reliable as the tests stated in the paper would 
indicate, it will come into extensive use in the near future, more 
especially as this would enable the engineer to use spans of several 
hundred feet greater length than is possible at present. The writer, 
however, has not checked Mr. Waddell's figures, on which the cost of 
long spans was compared, but it would seem doubtful if it would be 
possible to make an increase to the extent of 500 ft., mainly on 
accoimt of so many other factors than the mere cost of the metal 
entering into the cost of such structures. 

The data available for reference by the writer, in addition to this 
paper, would seem to indicate that the machining of nickel steel would 
be much more diflScult than that of ordinary carbon steel, although, 
with the small percentages discussed in the paper, for actual use this 
would not seem to be a very serious matter, and could only be 
determined definitely by actual experience in the shop. 

The tests also seem to indicate that the opinions held in the past 
as to nickel steel are in the main correct, and that the effect of the 
nickel is quite uniform. Would it not be well, however, to make a 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 301 

careful investigation of acid nickel steel, as there is no question that Mr. Fowler, 
for a high grade of steel for long spans, the acid process is some- 
what better than the basic. 

With reference to the future cost of nickel, it would hardly seem 
likely that even with the additional deposits which have been dis- 
covered, the cost would be very greatly reduced, and it would be well 
to know whether or not the author has investigated the use of fer- 
ruginous nickel for making the alloy, which would be just as effective, 
and of very much less cost. 

It will undoubtedly be necessary to make a great many additional 
tests before nickel steel can be put into general use, and it is to be 
hoped that some of the large steel manufacturers will undertake this 
work, which, to the necessary extent, can hardly be carried on at private 
expense. 

M. F. Brown, M. Am. Soc. C. E. (by letter) .^One of the most M"". Brown, 
important facts to be determined in connection with the use of nickel 
steel for bridges is the proper proportion of nickel and other elements 
to make the resultant alloy the most adaptable for the purpose intended. 
Br. Waddell is evidently of the same opinion, but it seems to the 
writer that his paper does not give enough prominence to this fact. 
The difficulties in the way of a private individual who may undertake 
to determine these proportions are apparent; and certainly few, if 
able, would be willing to devote the time and money necessary to such 
a solution, as this will hardly be found except after the lapse of 
years. It is probable that nickel steel, as a material for bridge spans 
of ordinary length, if it is ever commonly used for such purposes, will 
have to pass through some such period of development as has struc- 
tural steel during the last fifteen or twenty years, and be specified with 
proportions giving elastic limits ranging from those but slightly greater 
than carbon steel up to that as high as possible, consistent with fabrica- 
tion under refined shop methods. From these extremes will probably 
issue a generally accepted material adaptable to ordinary shop methods 
of manufacture. Whether or not this material will be similar to that 
proposed by the author is impossible to determine; but it seems to the 
writer that just because two melts of steel having practically identical 
proportions give a product of good quality which can be readily 
fabricated into structural members, it is a little hasty to conclude that 
"it is not likely that any great improvement in the characteristics of 
the future plate-and-shape steel, as compared with those of these melts, 
will be affected." 

The author's tests are well chosen to show the suitability of the 
material for the purpose intended, and certainly indicate the fitness of 
nickel steel as a material for bridge construction in spans of ordinary 
length, these being of the greatest interest to most engineers. Long 
spans, while of great interest in themselves, do not come within the 



302 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Brown, range of experience of the majority of engineers, and can be treated 
in a measure apart, being of such bulk that special material more 
particularly adaptable to them may be easily obtained if desired. 

The tests for resilience are somewhat surprising, the general im- 
pression being that nickel steel has considerably more resilience than 
ordinary carbon steel; but the difference is not enough to prejudice 
its use. The corrosion tests are interesting as showing some compara- 
tive measure, with steel, of the action of gases, etc. It would have been 
interesting if these tests could also have been compared on an 
iron basis. 

The writer is not disposed to agree with the author that the results 
of his investigation make it "practicable to write specifications for 
nickel-steel bridges which will possess the same strength, rigidity, and 
general excellence of design as the best carbon-steel bridges that are 
being built to-day." Anything like a general specification at the 
present time is attempting too much. The author's tests, valuable as 
they are in affording an indication of the use of the material, are too 
few to serve as a basis upon which to write a complete general specifi- 
cation. If, however, as seems probable, the author intends these 
specifications to serve as a guide only until further information is 
available, then they can serve a valuable purpose. As applied to the 
material as specified, the only criticism the writer offers concerns the 
compression formulas, which are based upon only six tests of full- 
sized members. This information appears to be too meager to 
develop a compression formula which will inspire confidence in the 
minds of conservative designers, and, until further tests are available, 
prudence would suggest reducing these units to an undoubtedly safe 
value. 

An enormous amount of labor is indicated in the preparation of 
the diagrams giving weights of bridges and their comparative costs 
with all carbon steel and mixed nickel and carbon steel. It seems 
almost a pity that the author should have gone into this so deeply 
when it is considered that the specifications for the two materials 
can hardly be expected to maintain the relations existing when the 
diagrams were prepared. However, they will illustrate their purpose, 
and the profession is in a position to appreciate them. 

Those remarks are not intended in a spirit of criticism of this most 
valuable and timely paper, and the writer wishes to add his word of 
appreciation of the spirit which makes such a paper possible and 
available to other and less gifted engineers. 
Mr. Bell. H. P. Bkll, M. Am. Soc. C. E. (by letter).— Dr. Waddell's paper 
is likely to be of great service, not only to the profession, but also to 
the general public. 

When it is considered that manufacturers can produce steel wire 
with an ultimate tensile strength of 100 tons per sq. in., and plate- 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 303 

and-shape steel of about one-third the same tenacity, it is evident that Mr. Beii, 
the makers of the latter should be encouraged to improve its quality as 
much as possible. 

It is a common expression that "it is the last straw that breaks the 
camel's back," and this aphorism comes forcibly home to the engineer 
who is designing a long-span bridge for modern double-track railway 
and other kinds of traffic combined. The use of nickel steel, as sug- 
gested by Dr. Waddell, may solve some problems not otherwise easy 
of solution. 

As soon as engineers begin to manifest practically a desire to have 
a superior quality of nickel steel for bridge purposes, the makers them- 
selves will take up the subject in earnest. The demand must come, 
before the quality of the supply will be as good as it can be made, and 
it probably will take some time to find out how to manufacture a 
superior kind of nickel steel for structural purposes. It is not to be 
supposed that an improvement once begun will be suddenly arrested. 
Upon the face of the facts already ascertained, there is something 
puzzling about the making of good nickel steel for bridge purposes- — 
something that requires continued action among the makers — to throw 
light upon apparent inconsistencies, and to clear the way to the per- 
fection that is likely to be attainable with due persistence. 

Not many years ago metallurgists found great difficulty in the 
reduction of refractory ores, but, being persistent in their efforts, they 
finally succeeded to a very large extent, but this improvement may 
continue for a long time still. 

"Resistance to cracking, a property to which the name of non- 
fissibility has been given, is shown more remarkably as the percentage 
of nickel increases. Bars of 2Y% nickel illustrate this property. A 
IJ-in. square bar was nicked J in. deep and bent double on itself with- 
out further fracture than the splintering off, as it were, of the nicked 
portion. Sudden failure or rupture of this steel would be impossible; 
it seems to possess the toughness of rawhide with the strength of steel. 
With this percentage of nickel the steel is practically non-corrodible 
and non-magnetic. The resistance to cracking shown by the lower 
percentages of nickel steel is best illustrated in the many trials of 
nickel-steel armor."* 

In a table on the page from which the above quotation is taken, 
there is the record of two specimens of Si% nickel steel (bars forged 
down from a 6-in. ingot to § in. diameter, with conical heads for hold- 
ing), the first with an ultimate tensile strength of 276 800 lb. and the 
second with 246 595 lb. per sq. in. On looking at the reduction of area 
and elongation, it is seen that neither rises above 6 per cent. 

As these ingots were not rolled but forged down, it looks as if the 
amount of work put upon them, while increasing the tensile strength, 

* See "The Mechanical Engineers' Pocket-Book," by Williain Kent, p. 408. 



304 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Bell, had impaired the reduction and elongation. It is in questions of this 
kind that manufacturers can be helpful to engineers, and it seems to 
the writer that engineers in specifying what they require, should not 
fall below the standards of this paper, but rather, if anything, exceed 
them, because the latter policy would put the manufacturers on 
their mettle. If it be assumed, in the first instance, that nothing better 
can be got than that which has been made already, progress and 
improvement will be slow. Suppose that a large bridge was to be 
tendered for and a notice such as the following was given : 

Every contractor must furnish with his tender bars of nickel 

steel of such and such size for testing purposes; and other things 
being equal, the tender possessing the strongest steel will have the best 
chance of securing the contract, which will not necessarily be let to 
the lowest bidder, but the contractor who is awarded the work shall 
bind himself to manufacture steel of at least equal quality with the 
samples tested. 

If such a notice (or one more carefully drawn) were used, the 
probability is that a superior class of nickel steel could be had in the 
shortest time possible. 

It was not to be expected that a preliminary inquiry, however ex- 
haustive, into a subject of such magnitude and importance to the 
profession and the public, would cover all the ground necessary to be 
gone over. This is a matter of time and probably also of expenditure, 
large enough to awaken the sympathy, and engage the active assistance, 
of all those who may be commercially interested in the making of 
nickel steel for structural purposes. 

Lately the writer became aware of the fact that a three-hinged 
metal arch of 1 800 ft. span could be built of carbon steel for a metal 
weight of 35 000 lb, per ft. ; no unit stress on main members of more 
than 15 000 lb. per sq. in., either in tension or compression ; and upon 
other parts none greater than 18 000 lb. per sq. in. ; width of the floor. 
80 ft. ; live load per foot, for double-track railway and other kinds of 
traffic, 16 000 lb. 

If this same structure were to be built of equal strength of nickel 
steel, it would weigh very approximately three-fourths of the above, 
or about 26 250 lb., instead of 35 000 lb. per ft. 

It will be evident to those who study Dr. Waddell's paper, that 
engineers could now safely build upon his specifications, and realize 
an important economy; and that as the improvement in the manu- 
facture of nickel steel continued, a class of material, better than the 
manufacturers would probably admit of as practicable at the present 
time, would soon be secured. 

The vn-iter believes that Dr. Waddell and his associates have done 
a most meritorious piece of work at an opportune time. It was need- 
ful that someone should take up the subject, and it is doubtful if it 
could have fallen into hands more capable, generous, and painstaking. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 305 

L. J. Le Conte^ M. Am. See. C. E. (by letter). — Experiments on aMr. LeConte. 
small scale to determine the unit stress per square inch on the raw 
material of new metals as they come from the mills is certainly highly 
desirable information to be used subsequently in the manufacture of 
bridge members; but this, strictly speaking, is not the material with 
which the bridge engineer has to deal. He is practically compelled to 
take the finished members as they come to him from the shops ready 
for erection. This is what he uses, and, therefore, the tests of full- 
sized members just as they come from the shops give him practical 
results of the highest importance, the intrinsic value of which cannot 
be over-estimated. Past experiments generally show "crippling" ap- 
parently somewhat inside of the so-called elastic limit. This, it seems, 
is undoubtedly due to the imperfections of detailing in shopwork which 
cannot be entirely avoided in any case. Shopwork detailing 20 years 
ago was very different from that of the present day, and, moreover, 
even to-day the work turned out under ordinary conditions may be 
vastly different from that turned out under a rush order. In the latter 
case the inspectors are often compelled to slur over many things which 
they would not think of passing otherwise, and the result is inferior 
shopwork, which reduces not only the strength, but also the life, of the 
whole structure. This is particularly true of built-up members requir- 
ing a large amount of riveting. 

This being the case, it is more than interesting to study the author's 
experiments on well-built, full-sized members of nickel steel. These 
columns he speaks of as properly designed and properly manufactured. 
The results reported are certainly highly satisfactory, and are to be 
commended in every respect. These experiments seem to show that 
the engineer is now called upon to make his unit stresses a function 
of the shopwork, instead of a function of the so-called elastic limit. 
In the West, particularly where the matter of railroad transportation 
of finished members is a very serious item of expense, it often happens 
that it is much cheaper to order the different parts of bridge members 
all carefully matched and punched at the eastern shops. Then, after 
their arrival at their destination, they are assembled and riveted up 
by the western shops. As a result, there is generally a mixture of good 
and poor work, and no formula could come anywhere near fitting the 
facts in such cases. 

Inasmuch as these experiments seem to show conclusively that the 
ultimate stress which finished bridge members will stand, when tested 
to destruction, is largely, if not entirely, a direct function of the 
quality of shopwork, it would seem to be entirely reasonable and proper 
that every important member of a bridge (tension as well as compres- 
sion — or in some members both) should be tested systematically with 
the full loads called for in the specifications before they leave the shop?. 
This requirement, although severe, would certainly develop the weak 
points in shopwork if any existed. Again, on the arrival of the struc- 



306 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Le Conte. tural iron at its destination, all members should be examined care- 
fully to see that they have not been injured in transit. And finally, 
during erection, care should be taken that the members are not injured 
in handling by the erecting parties. It is an old saying, and a good 
one, that eternal vigilance is the price of good work. 

The more the vpriter thinks over this paper, the more he is impressed 
with the fact that it is highly inadvisable to lower the standard unit 
stresses called for in all the standard specifications of the present day, 
because they are based primarily on long, painstaking, and wide- 
spread experience extending over many years of gradual and progressive 
improvement in shopwork. Furthermore, any attempt at reducing 
these unit stresses would be practically equivalent to offering a 
premium on cheap draftsmen and poor shopwork, both of which no 
engineer of experience in such matters would countenance. 

Whenever the quality of shopwork for any reason begins to de- 
teriorate, it is often extremely difficult to find out who is responsible 
for the unhappy results. Of course, true economy is always in order, 
and is certainly entitled to the highest respect, but true economy is 
also about the rarest article to find on the face of the earth. This 
arises from the fact that those in supreme charge rarely have any 
conception of the value of minute technical knowledge and experience 
in detailing shopwork, and very often discharge a capable and trust- 
worthy man in order to make room for a cheaper substitute. This is 
too often the case, and, as a result, they never seem to wake up or to 
realize fully the facts of the situation until some heavy bridge mem- 
ber, being handled by a yard derrick, buckles up right before their 
eyes. In shopwork, especially, the honest laborer, the world over, is 
always worthy of his hire. 

This paper is extremely instructive, and shows that the author is 
a master in the details of his line of business. It is written in a most 
fascinating way, and the happy results reported are so alluring that 
one is naturally and unavoidably drawn toward his side of the case. 
He shows conclusively the great economy in the use of nickel steel for 
bridge work in general. 

The enormous increase in the available length of maximum span — 
some 500 to 600 ft. — speaks volumes for all future designs in nickel- 
steel bridgework and structural materials of all kinds. Moreover, 
nearly 20% reduction in the cost of long spans will certainly put new 
life into the whole business, and will produce a great and lasting public 
benefit, the full measure of which cannot be comprehended by the 
average man. 
Mr. Hatt. W. K. Hatt, Assoc. M. Am. Soc. C. E. (by letter). — It may seem 
anomalous that the nickel steel tested by the writer for Mr. Waddell 
should show a greater rupture-work (resilience) than the carbon steel. 



DISCUSSION ON NICKEL STEEL I'OR BHIDGES 



307 



which is more ductile. The values quoted by Mr. Waddell are as Mr. Hatt. 
given in Table 61 :* 

TABLE 61.— Tensile Impact Tests. 





Number 
tested. 


Maxi- 
mum. 


Carbon. 

Mini- 
mum. 


A-ver- 
age. 


Number 
tested. 


Nickel (3.5o/o). 




Maxi- 
mum. 


Mini- 
mum. 


Aver- 
age. 


Elongation, 

Rupture- work 


5 
5 


32.0 
1910 


31.0 
1540 


31.5 

1736 


4 
3 


19.00 
2 300 


13.00 
1960 


16.5 
2198 



The writer has consulted a paper of his, entitled "Tensile Impact 

Tests of Metals,"t and finds other comparative tensile impact tests of 
nickel and common soft machine steel, as shown in Table 62 : 

TABLE 62. 





Number 
tested. 


Machine Steel. 


Number 
tested. 


3.15%' Nickel Steel. 




Maxi- 
mum. 


Mini- 
mum. 


Aver- 
age. 


Maxi- 
mum. 


Mini- Aver- 
mum. age. 


Elongation in 8 in. 
Rupture-work 


13 


31.6 
1460 


23.10 
1208 


27.00 
1S58 


8 


29.40 
2 216 


20.70 24.00 
1203 1821 



These tests, made in 1900, indicate the same phenomena as those 
made in 1907. In this same paper are tests on three grades of steel 
castings (coupons from locomotive driving-wheel centers). See Table 
64. These three tests would serve to indicate that the harder metal, 
with less elongation, may take a higher drop of a hammer for rupture 
than a softer metal of greater ductility. 

TABLE 63. — Impact Eupture-Work of Various Materlvls. 
Tension Impact with Single Blow. 



Material. 


Diameter, in 
inches. 


Gauge 
length. 

mches. 


Impact. 

Elonga- 
tion. 

Percent- 
age. 


Impact. 
Rupture 
work, 
in foot- 
pounds 
per cubic 
inch. 


Static. 

Tensile 

strength. 

in pounds 

per square 

inch. 


Soft stet-l 

Boilei- plate 


0.50 
1 by 0.5 ( rectangle) 


8 

8 


27.00 
34.40 


1 358 

1 855 


68 000 
60 000 




0.50 2 
0.50 1 8 


.33.00 2 315 
24.00 1 1 821 
0.70 186 
9.80 772 
13.60 765 
5.10 .556 


62 000 


Nickel steel 


85 000 


Steel wire.. . . .... 


0.16 108 
0.16 108 
0.30 lOH 
0.2(1 108 


115000 


Steel \vire annealed 

Steel wire annealed 

Steel wire 


83 000 
71800 
109 000 















*The writer has remeasured the drum records for these impact tests, and finds 
a correction to be made in the values of the rupture-work in the case of the nickel 
steel. The correct values are those above. 

t Proceedings, Am. Soc. for Testing Materials, Vol. IV, 1904. 



o08 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Hatt. This is not always the case, however, as may be seen from Table 63, 

which is arranged from the publication referred to. 

The elongation quoted is that remaining in the broken test bars 
after the impact test. At the exact time of rupture the elongation is 
undoubtedly greater by the amount of the subsequent elastic recoil, 
which latter is no doubt a function of the hardness of the steel. 

TABLE 64. — Tension Tests in Impact on Steel Castings. 



Character of fracture. 



Silky 

Flaky 

Fine bright 



Kind of test. 



Impact. 
Impact. 
Impact. 



Elongation. 



30.0 
19.0 
22.6 



Rupture-work. 



2 160 

900 

2 811 



Contraction. 



40.0 
18.8 
23.2 



Mr. Ostrup. 



John C. Ostrup, M. Am. Soc. C. E. (by letter). — This paper is 
certainly very exhaustive in its scope, and it is no wonder that so 
much time was consumed in its preparation, and in making the 
numerous tests and calculations involved. The information and deduc- 
tions resulting therefrom are well nigh incalculable in value to the 
profession in general and to the bridge engineer in particular. 

Without any outside assistance, it is manifestly impossible for 
many individual engineers to carry on costly experiments on such a 
broad scale, and it is more than doubtful whether any steel manu- 
facturer would care to do so. 

It is also improbable that the steel producer will look upon the 
author's conclusions with a kindly eye; that is, commercially speaking, 
it is in the interest of the manufacturers to oppose any innovation 
which will require a refitting, either partially or totally, of his furnaces, 
his shops, tools, etc., without the full assurance of a commensurate in- 
crease, and consequent profit, in the use of the new alloy. This again 
is doubtful, for some time at least, for the author has shown that 
superiority in economy does not become a decided factor except for the 
longer spans, or when the price of carbon steel is high. 

However, the work has been completed, the profession has received 
the benefit, and the writer thinks that too much commendation cannot 
be extended to the author or to his associates in this undertaking. 

Taking up various points in the paper, the first refers to the impact 
tests made to determine resilience. 

The author here correctly states, in regard to this method: 
"* * * the total amount of abuse given to the metal to be a measure 
of its toughness and not of its resilience." 

This is quite true, inasmuch as an impact test, no matter how made, 
can only give comparative ideas as to the toughness and ductility of 
various metals, and not their exact resilience. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 309 

The resilience of a structural material is determined otherwise, Mr. Ostrup. 
viz., by its ability to absorb potential energy, under a load or stress, 
and to restore the same when the cause has been removed. In other 
words, below the elastic limit, resilience is a measure of ability in 
materials to perform work. 

When such load or stress is being gradually applied, the resilience, 
per unit volume, is expressed as follows : 

Direct tension, or ) TJoaiiionno = '-' 

2 E 



( Direct tension, or > Kesilience , 

l direct compression [ 



For a beam of rectangular cross-section, such as was used in the 
experiments, this formula reduces to: 

(BendinjT) Resilience.... = ipr^, 

where S = Stress per square inch, at the elastic limit, uniformly dis- 
tributed over the entire area; 
S^ = Stress per square inch, at the elastic limit, in the extreme 

filiers ; 
E = Modulus of elasticity. 

Neither 8 nor S^ can be ascertained with any degree of accuracy, 
except where a quiescent load is being gradually applied. 

In a drop test, assume a weight, P, falling through a height, /i; then 
the total work = Ph. Part of this energy is absorbed by the supports, 
part by the polar moment of inertia of the beam, and the remainder 
goes to perform potential work. How great that remainder is, it is not 
possible to determine with accuracy. 

On the basis of the above equations, and using values specified in 
"De Pontibus," and also in this paper, a better comparison is obtained 
between the elastic resiliences of carbon steel and nickel steel, viz., 

40 000- ,_ . „ 

High-carbon stec4 = ^ xlo 000 000 ^ '*' "' P'^'' '""■ '"' 

Low-nickel steel = ., ^ 3Q ^^qq qqq = f><'''^'" '"-l'»- P^''' '■"• '"■ 

Iligh-nickel steel = -^r.r^T^ = 70.42 in-lb. per cu. in. 

2 X :50 000 000 

Or, calling the resilience for carbon steel 100%, we have: 

High-carbon steel 100% 

Low-nickel steel 216% 

lligh-nickel steel 254% 

These percentages would indicate correctly the comparative re- 
siliences were it not for the facts, as mentioned hereafter, which some- 
what modify the same. 



310 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. ostiup. When thus modified, it will be found that the theoretical com- 
parisons will approximate closely to those found in actual tests by 
Professor Hatt, as referred to by the author. 

Is it not quite true that the above coniijarative values, or in fact 
any exact comparison, will only obtain where perfect specimens — 
i. e., specimens uniform in cross-section throughout — are being tested. 
By the introduction into the tests of deformed or notched bars an 
element causing great uncertainty has been added. 

Compare, for instance, the flexural resistance of two bars, one of 
which is of somewhat larger cross-section than the other, but has been 
notched so that its section modulus (but not its moment of inertia) 
through the notched portion equals that of the smaller and uniform bar. 

According to theory, the two bars ought to carry the same load, 
but this is not so. The efiect of the notchings, by contracting the area 
of the cross-section, is to set up severe secondary stresses, their magni- 
tude being dependent on the relative size and shape of the notches with 
reference to the un-notched cross-section. 

It has been the writer's experience that these secondary stresses 
reduce the ordinary strength at the elastic limit by from 20 to 40%, 
and at the ultimate limit slightly more. 

While this gives an idea as to the reduction in the static strength 
of a notched beam, it shows by no means the reduction in its total 
resilience. This reduction follows an entirely different law, and the 
effect of this law is to concentrate a large part of the entire work 
(done upon the beam) on the weak section. 

This law is expressed as follows, when considering an elementary 

area: 

, Mcl/3 

dw = — ^ 

where w = Internal work done, 
M = Bending moment, 

13 = Angle after bending between planes which were parallel 
before bending. 

To those who have broken notched specimens, whether in the test- 
ing machine or across their knees, it is a matter of knowledge that 
the angle, jS, will be far greater at the point where the cross-section 
is reduced, hence the work done at -that point will increase and the 
total resilience of the section decrease correspondingly. 

That the author was aware of these facts, he indicated by finally 
intending to use plain bars, and it is the writer's opinion that all 
tests should be made on perfect specimens and, furthermore, that the 
"impact" test of any kind of steel be entirely dispensed with. 

The results obtained by the corrosion tests, and the lesson these 
contain, are certainly very instructive, if not somewhat ludicrous. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 311 

Tlius, for iustaiicc, in the acid lest the aiillidr's iiij;li-iiickel steel lost Mr. Ostrup. 
94% in IGO days, whereas that tested by The Osborn Engineering Com- 
pany lost apparently in the same space of time only about 2 per cent. 

This, however, is not mentioned as a reflection upon the tests 
themselves, but rather upon the fact that, so far as the writer knows, 
the profession has no standard specifications covering the method of 
performing either corrosion tests, or many other equally important 
tests, upon structural materials. 

It has been the writer's experience, in this matter, that nearly every 
testing bureau or testing laboratory uses a method of its own. 

In his specifications for nickel steel, the author advocates its use 
for the floor systems of bridges, saying "* * * the floor system 
shall preferably be of nickel steel." To this the writer agrees heartily, 
inasmuch as there should result a considerable saving in the web by 
its use. 

The economical depth of a stringer, or any plate girder, being 
dependent upon the allowable unit stress in the flanges, a much 
shallower depth with nickel steel could be used. Then again the 
deeper the web (where no intermediate stifieners are used) the thicker 
the web must be, in order to satisfy the requirement for a reasonable 
ratio between unsupported depth and thickness. 

In other words, with an allowable ratio of 1 : 60, a web plate 40 
in. deep, using 6 by 6-in. angles, would require a thickness of i in., 
whereas with nickel steel, for the same case, the depth of the web 
would only be 31 in., and its consequent thickness § in. Since there 
is nearly always a great surplus of metal in the web, the waste would 
be less when using nickel steel. 

In his comparisons of costs between carbon steel and nickel steel, 
with reference to plate girders, it would be interesting to know 
whether or not the author actually used the reduced economical depths 
allowed by the use of the stronger alloy. 

In many standard specifications giving working stresses for mem- 
bers under a direct stress, no mention or provision is made for stresses 
due to bending from their own weight, or for other secondary causes. 
This omission is more general where tension members are concerned. 
That such bending stresses are by no means negligible can be readily 
seen by reference to a recent case, where the sectional area of some 
members ranged between 700 and 800 sq. in., and the lengths between 
50 and 60 ft. 

In "De Pontibus" the author provides for such bending stresses, 
and whenever done, the use of nickel steel again will show an advan- 
tage, inasmuch as these stresses will be considerably reduced. 

The author does not specifically mention suspension bridges, where 
he demonstrates the economy in the use of nickel steel and its superi- 
ority for long spans, but such bridges should undoubtedly be included. 



313 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. ostrup. The case is nearly parallel to that of cantilever bridges, since its dead 
load is great compared to the carrying capacity, and any reduction in 
the dead load is also here of the greatest importance. 

Furthermore, the use of nickel steel in tall office buildings should 
be considered in the near future. Some of these buildings in New 
York City contain from 25 000 to 30 000 tons of steelwork, and, whereas 
the plate girders or beams might show little or no economy in the 
use of nickel steel, the writer thinks it highly probable that the columns 
would show considerable economy in weight and cost, and also ad- 
vantages in a reduction of the space occupied. 

The author states as his opinion that, when used in large quantities, 
the price of nickel would probably be reduced from 30 cents to 
25 cents per pound. Is not that contrary to the laws of supply and 
demand? Would not the use of nickel steel, or nickel, in large quan- 
tities equally well cause an advance in its price? 

Considering in its entirety the whole question of the adoption for 
bridges — or more broadly speaking for structural purposes — of nickel 
steel, the writer cannot add his voice too strongly to those who are in 
favor of an alloy which shows so many patent advantages. 

There will undoubtedly be some who, for one reason or another, or 
for no reason at all, will protest against its introduction. Many of 
our professional forefathers did the same when Bessemer steel was 
first spoken of, claiming that they had no confidence in its strength 
or suitability for structural purposes. 

Mr. Fidier. T. Claxton Fidler, Esq., (by letter). — This paper conveys informa- 
tion which is of much value and interest, and it goes so thoroughly 
into the economic advantages of the new material that very little can 
be said on that head. 

The calculations of steel quantities for variovis spans are very 
complete, and, although the methods of calculation are not given, the 
writer cannot doubt but that the special conditions which must affect 
the design of all compression members, and also those which arise from 
the undetermined value of the dead load in long spans, have been 
taken into account. It is only on these points that he ventures to 
add a few words. 

The experiments seem to justify a safe working stress, /„, for 

nickel steel, which is nearly the working stress, fc, that would be 

adopted for carbon steel; and this ratio of 5 to 3 seems to be applied 
to eye-bars and also to columns of the ordinary proportions of length, 
I, to radius of gyration, r. It follows, apparently, that, with the 
substitution of nickel steel for carbon steel, the sectional area of an 

'\ . .") 

eye-bar might be reduced to about \, making An = \ Ag. In the 




DISCUSSION ON NICKEL STEEL FOR TUaDGES 313 

c-aso of ;i coluiiin ciiriving any j^ivdi load. Iiowever, ^ Mr. Kidler. 

It would not be found practical to effect the same 
reduction, because it would be difficult to preseive, 
in both materials, the same ratio of I to r. 

Let us suppose, for example, that, in a brid2:e 
of given outline and dimensions, we have to design 
a column of given length, L to carry a given load; 
and, first, using carbon steel, we have fixed upon 

the best diameter, cZ, radius of gyration, r, and thickness of plate, «, giving 
a sectional area, Ag, proportional to the product r t. Then if we 
proceed to substitute nickel steel for carbon steel and keep to the 

same radius, r, we should have to reduce the thickness to t„ = \ t^, 

o 

and the column would fail by secondary buckling or a crumpling of the 
thin plate which was already thin enough. 

On the other hand, if we keep to the same ratio of r to t, in order 
to avoid this weakness, we should have to reduce the radius to some- 
thing like r„ = r,,\/0-60 = 0.8 ?>, or thereabout, and, when the best 

dimensions are found, the ratio will be greater than before, and 

/• 

the area J.„ greater than 0.6 Ac- 

So far as the compression members are concerned, the economic 
advantages of the stronger material are only fully realized when the 
load is great in comparison with the length, I, or, in other words, 
when there is no liability to buckling; and the advantage is gradually 
lost as the load becomes less in comparison with the length of the 
strut. In bridges of the longest span the columns will be of great 
length, and will be specially liable to buckling; and here we should 
expect to find, at the same time, that the dead load would be far less 
in a bridge of nickel steel than in one of carbon steel. The economic 
advantage of the stronger material, as applied to these long and 
slender struts, would be comparatively small, just as it has been in 
the substitution of carbon steel for wrought iron ; and, though it could 
not disappear altogether, it would be less than the advantage of 3 to 5 
in the weight of steel required. 

On the other hand there is no such limitation in respect to the 
tension members, and if eye-bars of nickel steel are found to fulfill all 
the practical requirements, there seems to be no reason why they 
should not realize the economic advantage to the full. In suspension 
bridges we should get this advantage throughout the main super- 
structure; in girders or cantilevers we should get it in the tension 
members, which form one-half of the structure, while in arches the 
advantage would be still less. 

The writer has long believed that a suspension bridge of rigid 



314 DISCUSSION ON NICKEL STHEL l-Olt BRIDGES 

Mr. Fidler. construction would efficiently fulfill all conditions of railway traffic, 
while realizing other important advantages (a view which was sup- 
ported on one occasion by Lord Kelvin), and, for spans of 1 500 ft. and 
upward, it would possess (in many locations) a distinct advantage 
over the cantilever in point of economy as well as in its rigidity under 
traffic, its unfailing stability, and its safety under wind pressure. 

In any case it is evident that the economic advantage of nickel 
steel, whatever it may be for small spans, will be enormously greater 
for bridges of very long span where the weight of the steel itself forms 
the chief part of the load to be carried. The curves of the author's 
diagrams show very clearly how the weight of steel, w, per linear foot 
increases rapidly with the span, L, rising toward infinity as L 
approaches the "limiting span," L , for any given type of structure. 

In a discussion of this matter in "Bridge Construction," the writer 
has given the formula for weight of main superstructure per foot, 

'tn = — 

1 — i S7M 

while the total weight of the structure is w; + p. 

Here the wind metal is reckoned separately and included as part 
of the platform load, p, and therefore part of the load carried by the 
main superstructure. 




' , , 1 • 1 

Fig. 77. 

For the main superstructure, the square inches of steel section 
required in each member are separately reckoned for the three func- 
tions which have to be fulfilled, viz., 

For carrying the platform load the quantity L p'Siy fi' 

For carrying the live load and its impact L 7 2 7 m" 

For carrying its own weight L w "S, y fx 

The formula is applicable to independent girders, arches, or suspension 
bridges, but, as it stands, not to cantilever bridges; and for a given type 
of design the quantities, ii, /j.', and m", are independent of the material 
used. The limiting span, /^ , is obviously reacluMl when 1 — i S 7 ja = 0, or 
when i S7/ii = 1, so that^ is inversely proportional to the specific weight, 
7, iwv ton of stress, or directly projjortional to tlie working stress, /. 

For each type of structure in nickel steel, /,'^j = £^. *" ; and for the sus- 

Jr 
•~) 

pension bri^lge we should expect/;,, = ., £^,. 



DISCUSSION ON NICKEL STKEL FOR BRIDGES 315 

The curve representing the relation of w to L is easily traced ; and Mr. Fidier. 
its general form agrees with the curves of the author's diagrams, as 
shown in Fig. 77. 

Egbert E. JohiNSTon, Esq. (by letter).— The writer congratulates Mr. john.ston. 
the author on the results he has obtained in his investigation of the 
physical characteristics of nickel steel as compared with carbon steel. 

After a careful perusal of the paper, the writer is of the opinion 
that, with the proportions of nickel determined by the experiments, 
nickel steel has been proved to be a material which can be safely 
relied on in the construction of bridges of large span, and that con- 
siderable economy can be obtained by its adoption as compared with 
carbon steel; but it is a question whether the difference in cost will 
justify its adoption in bridges of small span, because, for structural 
reasons, full advantage cannot be taken of the difference in the 
strength of the two materials, for instance, in the webs of plate 
girders. 

The difference in the strength of nickel steel for rivets as com- 
pared with carbon steel, requires that, for the latter material, addi- 
tional metal must be provided in order to compensate for the area 
removed from the plate, owing to the increased diameter to be given 
to the nickel-steel rivets, as stated in the paper.* 

Another great advantage of nickel steel over carbon steel is that 
the former is affected to a less degree by the atmosphere and the 
fumes from locomotives. This is of great importance in the maintenance 
of large bridges, for it necessitates a considerable outlay in scaffolding 
for painting purposes, but, to a certain extent, this is discounted when 
the two metals are used in the same structure. 

The pliotograplis of the colimms tested (Plate XIX) would appear 
to indicate that the unsupported wing of the angles should be made 
thicker than the other wing, which would have the effect of increasing 
the radius of gyration, and thus add to the strength of the column. 
It would add to the interest of the experiments if they included cylindri- 
cal columns, in order to determine their strength and also to ascertain 
the ratio of the thickness to the diameter. 

Albert Lucius, M. Am. Soc. C. E. (by letter).- — The writer would Mr. Lucius, 
most assuredly build long-span bridges of nickel steel. The disclosures in 
reference to the Quelec Bridge and the Blackwell's Island Bridge 
would not leave a moment's hesitation in his mind. With equal 
assurance, he would build short-span bridges of soft steel, and would 
not even be influcMiccMl hy the i)o>siliility tliat niekel-steel bridges 
might be as cheap or cheaper, simply on grounds of greater perfection. 
The writer is not sure where he would draw the line, but, in a general 
way, he would draw it when dead-load stresses exceeded live-load 
stresses, and would then commence to study the matter in detail. 

* Page 116. 



thai. 



316 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Lucius. Dr. Waddell's paper is a complete-enough proof that nickel steel 

is fully available. The writer would prefer heavy bridges of soft 
steel to light bridges of nickel steel for railroad purposes, because 
he would consider them steadier under heavy loads and high speeds. 
When the practical limitations of manufacturing bridges are reached 
he would use nickel steel to help out, and would probably make the 
entire construction of nickel steel, and not only a few members. The 
\vriter considers this paper to be highly valuable, and believes that the 
thanks of the Profession are due to the author. 
Mr. Linden G. LiNDENTHAL, M. Am. Soc. C. E. (by letter). — From his experiments, 
the one conclusion which Mr. Waddell evidently desires to emphasize 
most is that nickel steel is economical, as compared with the usual 
structural carbon steel, for nearly all kinds of steel structures. Mr. 
Waddell's reasons for such a claim are ample and weighty, but not 
always applicable. It is also doubtful if the economy shown in his 
tables could be obtained if nickel were extensively used. A large 
demand for it would probably raise its price. Nickel is not a very 
abundant metal, nor is it very cheaply produced. 

The use of high steels, nickel steel being one of them, in bridge 
construction, is justified only in very long spans, meaning that length 
of span in which the dead load is considerably greater than the live 
load. The average stress in bridge members from live load is then 
smaller than that from dead load, and that keeps the deflections of the 
structure within, moderate limits. 

In shorter spans, when the stresses from live load exceed those 
from dead load, the structure becomes too springy, vibrations become 
greater, and the effect of impact, of which we know so little, becomes 
very pronounced. 

We recognize the relation between train and bridge to be that of 
hammer and anvil. The anvil would not last long if it were much 
lighter than the hammer. To use high steel (including nickel steel) 
in a light structure, for the sake of first low cost, is not true economy. 

Although nickel steel was proposed for bridge construction long 
ago,* the writer was the first, he believes, to introduce it practically 
in the form of forged eye-bars, namely those for the Blackwell's Island 
Cantilever Bridge over the East River in New York. The design for 
this bi'idgc originated with the writer (in 190.3), but he resigned from 
the charge of this work about one month after making the contract 
for it. lie had no conucetion with, and is not responsible for, the 
changes subsequently made, which resulted in an inferior structure. 
Among the changes attempted was the substitution of hard (high- 
carbon) steel eye-bars for those of nickel steel ; but the tests with them 
were failures, and so the eye-bars of nickel steel were retained, and 
these are the bars described in Mr. Waddell's paper. 

* "Suspension Bridges— A Study." I)v Genrge S, IMorisoii, Tnnisartiitiis. Am. Soc. C. K.. 
Vol. XXXVL p. 850. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 317 

To avoid the great loss of strength in forged eye-bars through Mr. Linden- 
annealing, it would be more economical to make large eye-bars by ^ '^ ' 
cutting them from plates having the width of the head, as was done 
for the eye-bars of carbon steel in the new Budapest Suspension 
Bridge. 

According to Mr. Waddell's tables, it would seem that cantilever 
bridges are the only ones worth considering for long spans; but that 
is not true, even from the purely commercial point of view, as there 
are other systems, as arches and continuous girders, which for long 
spans are more economical and more rigid. The cantilever systems 
mentioned in his tables are, of all types, just the ones in which nickel 
steel would increase their already sufficiently great defect, that is, 
cumulative vibration and lack of rigidity. 

Nickel steel will be found to be most appropriate and economical 
in the suspension type of bridge. The writer, in 1903, proposed that 
material for the eye-bar chains of the Manhattan Bridge over the 
East River; with spans of 725, 1 470, and 725 ft.* 

Had a competition of the nickel-steel chain design there described, 
viz., steel-wire cables design, not been suppressed by interests adverse 
to the public good, there is no doubt that the nickel-steel eye-bar bridge 
would have been found very much cheaper, besides being better in 
every other way. 

The wrong perception which is still prevailing among engineers in 
regard to the nature of suspension bridges of long span, and, when 
properly designed, their eminent safety and suitability for heavy and 
fast railroad trains, may be explained perhaps from the fact that such 
bridges can be built in only a very few localities, and therefore have 
been insufficiently studied. In the writer's opinion, nickel steel and 
other high steels are the most suitable materials for them. 

Henry S. Prichard, M. Am. Soc. C. E. (by letter). — Nickel-steel Mr. I'richard. 
eye-bars have been manufactured and used to an extent which justifies 
their classification as an article of commerce when ordered in sufficient 
quantity, but so meager has been the use of nickel steel for the riveted 
members of structures that it has hardly reached the experimental 
stage; rather it is a matter of academic discussion. 

It is possible, of course, to draw some conclusions from tests such 
as the author describes, but the most that can reasonably be expected 
from them is that they will help to determine whether or not it is 
worth while to try the experiment of riveted nickel-steel structural 
work. Even if preliminary investigation is favorable, nothing but 

* " General Methods for the Calculation of Statically Indeterminate Bridges, as used in 
the Check Calculations of Desierns for the Manhattan "Bridge and the BlacKwell"s Island 
Bridge. New York:." by Frank H. Cilley. Transactions. Am. Soc. C. E., Vol. LIII. p. 413. 

" A Rational Form of Stiffened Suspension Bridsre." by Gustav Lindenthal, .M. Am. Soc. 
C. E., Tiansactionn, Am. Soc. C. E.. Vol. LV, p. 1. 

"Theory and Formulas for the Analytical Computation of a Three-Span Suspension 
Bridge witn Braced Cable," by Leon Moisseiff, M. Am. Soc.'C. E.. Tra^isactions, Am. Soc. 
C. E.. Vol. LV, p. W. 



318 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Prichard. actual experience can show conclusively the suitability of the material 
and the cost of fabrication. 

The suitability of the material has to be gauged, not onlj^ by the 
ordinary treatment it receives in fabrication, erection, and service in 
structures, but, in addition, by the abuse it can stand and to which, in 
spite of reasonable care, it is more or less subject, both before and after 
the structures are completed. 

Existing data are not sufficient to admit of a close reliable estimate 
of the relative cost of nickel and ordinary steel riveted work, but the 
writer regards the author's estimate as too favorable to nickel steel. 

The position of engineers with regard to the use of nickel steel 
is similar to that of the girl whose mother expected her to learn to swim 
without going near the water; they should not use nickel steel until 
they have had experience with it, and they cannot get the experience 
until they use it. 

Those who desire to be pioneers in the use of nickel or other special 
steels, if the preliminary investigations are sufficiently favorable, 
should consider other matters in addition to the results of experiments 
and preliminary estimates of cost. For the ordinary run of structures, 
if structural steel of a special quality is desired, the quantity is com- 
paratively so small that it is difficult, even at an increased cost, to 
obtain it. When it is possible to do so, it would doubtless give greater 
security and longer life to such structures to put the increase in cost 
into more metal of standard quality. There is great virtue in body, 
and many structures could have more with advantage. 

Especial caution is necessary in using for ordinary structures unit 
stresses greater than would be safe in ordinary carbon structural steel. 
Even if the engineer is satisfied that he can rely on the superior 
strength of some special kind, he must reckon with the fact that such 
steel cannot be distinguished from the ordinary variety by simple 
observation, and that, unless the mills are making and tlio manu- 
facturer using the special steel exclusively, there is a likelihood that 
some steel of ordinary quality will get mixed with it, in spite of the 
best intentions and reasonable care. If he succeeds in getting only 
the special steel, there is danger that in course of time the structure 
will come imder the supervision of other engineers who will not know 
the quality of the steel, and who. in judging as to safety, will be 
obliged to assume that it is no better than ordinary. 

For bridges of very long span, the dead weight is such a large pro- 
portion of the entire load that the comparative lightness obtained by 
the use of steel of great strength is a decided advantage, osixH'ially 
for tension members. IVTembers in compression, in addition to resisting 
the effort of the load to crush them, liave to resist its tendency to buckle 
and wrinkle them, and the resistance to these tendencies is about the 
same for steel of all grades. 



DISCUSSION ON MCKKI. STEEL FOR BRIDGES 319 

For bridges of very long span it is possible, and for the eye-bars Mr. Prichaid. 
of such spans it is probable, that the advantage of steel of great 
strength may offset the disadvantage of increased cost. There is, how- 
ever, great need for caution in regard to riveted members. 

For eye-bars there is, fortunately, a precedent in the Blackwell's 
Island Bridge, the full-sized tests of which are given in Table 60, 
which shows the results of 125 full-sized tests. The percentage of 
nickel varied from 3.22 to 3.76, with an average, according to the 
author, of 3.39. The percentage of carbon varied from 0.32 to 0.46, 
with an average, according to the author, of 0.39. The ultimate tensile 
strength per square inch varied from 80 360 lb. (16 by 2J-in. bar) to 
98 980 lb. (16 by 2|-in. bar), and the recorded elastic limit from 
39 960 lb. (16 by 2TV-in. bar) to 60.570 11). (16 by lf^in. bar). The 
percentage of elongation in 18 ft., omitting bars that broke in the head, 
varied from 8.44 to 18.44. In 27 of the 12.5 cases the recorded elastic 
limit is less than 55% of the ultimate strength, and in three cases it is 
less than 50 per cent. Owing to the large number of tests, the 
character of the plant which manufactured the bars and made the tests, 
and the meagerness of other available data, they should carry great 
weight with engineers in posting themselves as to what it is practicable 
to manufacture, and in avoiding what it is useless to specify. 

The author tested eight nickel-steel eye-bars which he states were 
almost identical in composition with those for the Blackwell's Island 
Bridge. The ultimate tensile strength per square inch varied from 
88 900 lb. (8 by 2-in. bar) to 103 200 lb. (6 by 1-in. bar), and the 
recorded elastic limit from 48 300 to 58 700 lb. per sq. in. The per- 
centage of elongation in 10 ft. varied from 10.6 to 14.5. In four of the 
eight cases the elastic limit was less than 55% of the ultimate strength, 
and in one case it was 51 per cent. Two of these four cases were 
6 by 1-in. bars, a size in which the best results are attainable. 

In addition to the tests cited, the author tested two eye-bars 
fabricated from f-in. plates, with planed edges, of 4.25% nickel and 
0.46% carbon. The results of these tests, naming number one first in 
each case, were : Ultimate intensity, 105 900 lb. and 102 300 lb. ; elastic 
limit, "Lost" and "Uncertain"; percentage of elongation in 10 ft., 
7.4 and 6.8. 

The author's speeifications for eye-bars call for: Nickel, 4 to 4.5%, 
and carbon 0.40 to 0.50% ; and require full-sized tests to show an 
ultimate tensile strength per square inch varying from 90 000 lb., for 
2^-in. or greater, to 105 000 lb. for 1-in. They require an elongation 
of not less than 10% in a gauged length of 10 ft., and an elastic 
limit of not less than 55% of the ultimate strength of the bar. These 
specifications are much higher than are warranted by the tests cited. 
With two exceptions, the tests have no direct bearing, as the steel had 
lower percentages of nickel and carbon, but, inferentially, they do not 



320 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Prifhard. support, but are ag-aiust, the autlior's specifications. Of e.ve-l)ars of 
about the quality specified, the author cites only two full-sized tests, 
neither of which is up to the specifications. In discussing these two 
tests, it should be remembered that the edges of the bars were planed, 
that the manganese was less than the specifications require, and that 
the author is of the opinion that rolled edges and manganese as per 
specifications would have produced better results. Until he gets the 
desired results, however, and in sufficient quantity to show that it is 
practicable to rely on them in practice, his specification for eye-bars is 
not warranted. 

The author's specification of 55% for the elastic limit is in harmony 
with his statement on page 139, under "Full-Sized Eye-Bars."' that 
"The elastic limit in nickel steel never falls below 55% of the ultimate 
strength," but it does not agree with the tests of the eye-bars for 
the Blackwell's Island Bridge, nor with his own full-sized eye-lar 
tests, for half of which (four out of eight) the recorded elastic limit 
is less than 55% of the ultimate strength, with a minimum of 51 per 
cent. It should be noted here that the specifications require that: 
"the elastic limit shall not be less than 55% of the ultimate strength 
of the bar," instead of limiting it to 55% of the ultimate strength 
required. The difference can be illustrated by reference to two eye-bars 
2 in. thick, the first with an ultimate of 95 000 lb. (the amount re- 
quired) and an elastic limit of 52 250 lb., and the second (an actual 
bar) with an ultimate of 95 880 lb. and an elastic limit of 52 360 lb. 
The first bar conforms to the author's specifications, but the second 
bar (of the same size) does not, although it has a higher elastic limit, 
because this limit is less than 55% of the ultimate strength of the bar. 

Structural steel which has not previously been subjected to an 
external load usually displays slight imperfections in elasticity, and 
acquires a slight permanent set before, and sometimes long before, 
there is any considerable loss in elasticity. Eventually, under a 
gradually increasing load, if the steel is moderately soft, the yield 
point is reached where the low rate of change, characteristic of steel 
strained within its elastic limit, suddenly changes to the high rate 
which results from the flow of the metal after the destruction of its 
elastic properties. As the percentage of carbon is increased, the abrupt 
change in the rate of extension becomes less marked and finally dis- 
appears, so that there is for such cases no sharply defined yield point. 
The attempts to define it clearly for hard steel are arbitrary, nud wIkmi 
rigidly adhered to may be unjust and unwise. 

The elasticity of steel within certain limits is made nearly, perhaps 
quite, perfect by a second or subsequent application of the load after 
a suitable interval of rest, but the material has to take some perma- 
nent set in order to reach this condition. As a basis for designing 
steel structures, the critical question with regard to strength is: "What 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 321 

is the greatest permanent or indefinitely repeated load which structural Mr. Prichard. 
steel will sustain without undue deformation?" It 'would be very 
satisfactory to have this question very closely settled in each case by 
some direct and simple test. Unfortunately, this is not possible. 
Judging from the tests and experiments which have been made, it is 
not wise to count on an elastic limit for eye-bars of more than 50% of 
the ultimate strength, or to expect that there will not be slight imper- 
fections in elasticity below this limit on the first application of the 
load. 

The author quotes from a paper by the English authors, Messrs. 
Carpenter, Hadfield, and Longmuir, to the effect that steel containing 
from 0.40 to 0.50% of carbon and 4|% of nickel has lost none of its 
toughness or ductility, but that a steel containing 5% of nickel has 
lost both to a serious extent. He is of the opinion that between these 
two percentages there is probably a well-defined point of demarcation 
which can only be determined by further experiments. Would it not be 
well to wait for more information on this point before specifying eye- 
bar steel with nickel from 4 to 4i per cent. ? 

The author's specifications allow 66§% greater tension in nickel- 
steel eye-bars than in those of ordinary carbon steel. In the writer's 
judgment, this is not warranted by experiments, and is too great an 
increase, especially when applied to unit stresses as high as those 
quoted from "De Pontibus." The calculated stresses are only approxi- 
mations, and are not always close or complete; besides which, the 
likelihood of defects, the possibility of blunders, the certainty of 
deterioration, and the chance of accidents and unforeseen contingen- 
cies, make it unwise, to say the least, to narrow the margin of safety. 

The matter of annealing eye-bars, as the author states, is worthy 
of further investigation. A complete understanding of the phenomena 
involves a knowledge of the exact nattire of steel and the reasons for 
its properties. The nature of steel is, at present, largely a matter of 
conjecture, and therefore, it is not yet possible to make a satisfactory 
explanation of annealing. The prominent conditions and results in 
any given case can be observed, and, of course, the inference can be 
drawn that similar conditions will produce similar results; the results, 
however, can be decidedly affected by many things, some of which are 
quite subtle, both before and after the metal reaches the annealing 
furnace. 

Eye-bar manufacturers do, of course, make earnest study of the 
phenomena of annealing, and reach well-st;pported conclusions. In. 
the case of tlie nickel-steel eye-bars for the Elackwell's Island Bridge, 
the American Bridge Company made more than 250 full-sized tests. 
A very large proportion of these were made in advance of commercial 
manufacture, to study the annealing processes and temperatures neces- 
sary to produce satisfactory results. On page 13Y the author seems 



323 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Prichard. to question the care with which the eye-bars for this bridge were 
annealed. As he states, he was not furnished with information con- 
cerning the annealing, nor was he furnished with records of the 
experimental tests which, from the nature of the case, were really 
much more instructive than the tests on which the bars were accepted, 
and which he reported in Appendix B. 

The range of experimental tests on annealing made by the American 
Bridge Company was very much wider than that contemplated by the 
author, as outlined in the last paragraph of page 134 in variety of 
sizes, number of pieces of full length, number of pieces annealed at 
one time, etc. Temperatures, also, were varied within comparatively 
wide limits. These experimental tests established the necessity of 
treating each bar in accordance with its own particular chemical and 
physical characteristics; and this experience was applied with great 
care to the commercial product in a well-equipped and thoroughly 
controlled furnace. 

The opportunities of eye-bar manufacturers for observation are so 
great that their advice in the various cases which arise should be very 
valuable. 

Experience admits of some general conclusions regarding the an- 
nealing of eye-bars, but it is not wise, in the present state of knowl- 
edge, to formulate the process rigidly. 

Eye-bars are not annealed for the purpose of improving the quality 
of the material as it comes from the rolls, but simply to offset the 
undesirable results of forging the heads. In annealing eye-bars, the 
metal is softened, and some of the good effect of rolling is taken away. 
The end which should be sought is to offset the possible ill effects of 
forging the heads, with as little change as possible in the condition of 
the metal as it comes from the rolls, and this end should be kept in 
mind when applying the well-known fact that eye-bars which cool 
slowly will be softer than those which cool more rapidly. 

The thanks of the Engineering Profession are due to the author for 
the information which he has contributed in this paper, and he should 
be commended for the earnest efforts and generous expenditure which 
made this contribution possible. In endeavoring to establish closely, 
by tests and analysis, in advance of experience, the engineering worth 
and economic value of nickel steel as a structural material, his aspira- 
tions were very high, and he set for himself a task which few engineers 
would care to attempt. Wliile, owing to practically insurmountable 
difficulties, he has not accomplished all he desired, he has succeeded in 
increasing greatly the general knowledge regarding this material. 

Mr. Le^Chate- Henry Le Chatelier, Esq. (by letter).— The advantage of using 
steel of high resistance in the construction of long-span bridges is too 
evident to require any emphasis. Attention has already been called 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 333 

to this subject by Mr. Louis Le Chatelier.* After a series of breaking Mr. LeChate- 

tests on beams made of silicon steel, having an ultimate strength of 

101 000 lb. per sq. in., he first used this steel in an overhead crane 

girder, with a reduction of 25% in the weight of metal, and then in 

the derrick of an erection traveler where the weight reduction was 

still greater. 

Unquestionably" the same advantages can be obtained with nickel 
steel as with silicon steel; the choice of one or the other of these 
metals will be essentially a question of net cost. It would be of 
great interest to test these special metals systematically, and on a 
large scale, in heavy engineering constructions. Even if the cost 
should be greater than Mr. Waddell actually admits, it would still be 
advantageous to make these trials, at least by such agencies as railroad 
companies, or State departments, which often have to renew construc- 
tions of this kind. Even should the cost of these trials be excessive, 
they would still, in the end, result in structural advantages by which 
the above-named agencies -would be the first to benefit. 

It is doubtful, in the nieantimo, whether the use of these special 
steels will become general, unless there is a revolution in the net 
cost of the raw materials. It is very probable that, after a short-lived 
use of special steels, engineers will return to the ordinary carbon 
steels, but with the latter very much harder than they are now, just 
as was the case with the construction of torpedo-boats. 

On the advice of the French constructor Normand, there were 
adoi)ted, some twelve years ago, for the hulls of torpedo-boats, plates 
made of nickel steel containing 3i% nickel, and a proportionally high 
amount of carbon, altogether similar to the steel treated by Mr. 
Waddell; but very soon the steel manufacturers offered carbon-steel 
plates possessing the same resisting qualities, at a less cost, and 
it will certainly be the same for large metallic structures. Neither 
nickel nor silicon adds to the steel any special qualities of resistance, 
but these qualities are obtained with less care in manufacture than is 
required to produce a steel that is high in carbon and of correspond- 
ingly high strength, but without the fault of brittleness. It is much 
more difficult, and requires much more care in manufacture, to 
obtain these results with steel containing carbon only, but it is not 
impossible. The use of nickel steel must only be considered as a step, 
certainly a very useful one, and perhaps indispensable, in the advance 
toward the final use of hard carbon steel. 

In the desire to pass directly from the present soft steel to hard 
steel, there is danger of errors of application which may be of such 
a nature as to retard for a long time the definite use of steel of high 
resistance. The work of Mr. Waddell, therefore, cannot be too highly 
encouraged. 

* La Revue de M^talhirgie, 1904, p. 646. 



324 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Ross. A. Ross, EsQ. (by letter).- — Dr. Waddell's paper comes at an oppor- 
tune time, inasmuch as hitherto information on this subject has con- 
sisted of little more than detached paragraphs in the press; and this 
paper concentrates the past and gives valuable information and experi- 
ments for the first time. The subject is ripe for consideration, inas- 
much as engineers have been studying for some time the value of 
nickel steel. 

In Great Britain few structures, such as bridges, have been con- 
structed of nickel steel. Any experience the writer has had has been 
solely in connection with rails, having experimented on two or three 
occasions with steel containing, in one case, 3.3% of nickel, and, in 
the other, 1.5 per cent. In both cases the physical results obtained 
were better than with ordinary steel; but, looking at the matter com- 
paratively, and taking price and wear together, nickel steel did not 
show to much greater advantage, commercially; in some cases the 
nickel had not been uniformly distributed, in other words, the metal 
was not altogether homogeneous, and the writer came to the conclu- 
sion that the manufacture was rather more to blame than the 
analysis. 

Undoubtedly, it would be a great advantage in bridge building if 
one could secure a reliable alloy, which would bear higher stresses 
than ordinary carbon steel and, consequently, would be of relatively 
less weight. As far as. the writer knows, steel containing a percentage 
of nickel is the only metal of this kind we can look for. 

The writer observes that the nickel steel with which Dr. Waddell 
experimented was manufactured exclusively by the "basic open-hearth 
process", which he understands is the process adopted almost every- 
where in the United States, and which the author assumes will be 
that absolutely used in the future; but, in Great Britain, the acid 
open-hearth process is more generally used than the basic open-hearth. 
At the same time, nickel-steel rails can be manufactured with facility 
by the Bessemer process, which is that largely used in Great Britain 
for steel rails. 

With the little experience the writer has had with steel containing 
a percentage of nickel, he can easily assent to the valuable statement 
made by the author that for plates and shapes, which form the princi- 
pal part of bridge construction, any material with a greater percentage 
of nickel than 3.5 renders the alloy too refractory for the various shop 
manipulations to which it must be subjected in being manufactured 
into bridges. This fact, being established, fixes the limit of the 
percentage of nickel to be used to meet genernl requirements. 

As to carbon : It is pointed out that an addition of carbon might 
raise materially the ultimate strength and elastic limit of the metal, 
but this again is limited by the facility of workability in the shop, and 
clearly it would be more or less dangerous to have two classes of 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



325 



metal on the ground for the same purpose, viz., those which could go Mr. Ross, 
under manipulation in the shops and those which could not, as there 
would be difficulty in keeping them separate. 

The author's numerous tests, finally leading up to his conclusions, 
are very instructive, and comparing these with carbon steel, the safe 
limit of which, in Great Britain, is taken at 6^ tons, both in com- 
pression and tension, and at the standard of 1.00, the comparative 
results given by nickel steel are as shown in Table 65. 

TABLE 65. 





Tension. 


Compression. 


Bending. 


Bearing. 


Shear. 


Eye-bars 


1.67 
1.75 
1.71 


lies 

1.55 

1.48 


i!85 


1.73 
1.50 




Built members 


■ 

i 

1 
1 








Pins 


67 


Rivets 


40 




7(» 


Top chords of railway br 
Columns, fixed ends 


idges.. 




" hinged ends 











With the smaller members of a bridge it is difficult, even now, with 
carbon steel, to limit the sectional area of the bar to its absolute 
requirements, and it becomes more so with a metal of greater unit 
strength, such as nickel steel. The author's anticipitation of ordinary 
county bridges vanishing into thin air might be realized with dis- 
astrous effects on small railway bridges. 

The corrosion tests that were made do not appiear to be quite satis- 
factory or convincing, and, in the vpriter's opinion, they do not lead 
to the conclusion that ''nickel steel resists [corrosion] decidedly better 
than carbon steel," but leave the question open for further experiment 
and consideration. With equal resistance to loss of weight by corro- 
sion, a bar of nickel steel will deteriorate in strength 50 or 60% faster 
than a similar bar of carbon steel. 

Economical considerations will lead to the use of nickel-steel plates 
thinner than those of carbon steel. If the thickness is reduced rela- 
tively to the strength of the alloy, for. example, if a i-in. nickel-steel 
plate is used instead of a xV'iii- carbon-steel plate, then the percentage 
of the weight of bar lost by corrosion in the nickel steel will be one 
and one-half times as great as in the carbon steel. This increased loss 
of weight is equivalent to an increased loss of strength in the nickel 
steel at least two and one-fourth tinios as groat as in carliou steel. 
The introduction of nickel steel will involve, therefore, an increase in 
the present allowance for corrosion. 

At prices now current for nickel and for structural steel in Great 
Britain, the diagrams show that there is no economical reason for the 



326 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Ross, introduction of nickel steel for railway bridges of less than 260 ft. span 
in the case of double-line, or about 160 ft. span in the case of single- 
line bridges. 

Probably for a long time after the first introduction of this use for 
nickel steel, bridges of spans much exceeding the above limits would 
be built more cheaply of carbon steel, unless the price of nickel were 
to fall very much below its present rate of from £170 to £175 per ton. 

For small spans (30 ft.), the costs of nickel-steel, wrought-iron, and 
cast-iron bridges, compared with carbon-steel bridges, work out aa 
follows : 

Carbon steel 1.00 

Cast iron 1.07 

Wrought iron 1.20 

Nickel steel 1.23 

In Great Britain plate girders of ordinary carbon steel cost about 
3 cents per lb., or £13 16s. Od. per ton ; truss girders cost about 34 
cents per lb., or £16 2s. Od. per ton. If built of nickel steel, the 
increased price will not be less than 1.6 cents per lb., or £7 7s. 2d. per 
ton extra in each case. 

This gives a very large increased strength per unit, and a corre- 
sponding weight gives a very much less dead weight of metal required 
in the case of nickel steel as compared with carbon steel, clearly show- 
ing that an engineer can afFord to pay a greater price per ton for 
nickel steel than for carbon steel, but to what limit the writer is not 
prepared to say at present. 

No doubt many things are still required to be ascertained, and one 
feels that the manufacture of steel is always a delicate matter, and the 
more complex it becomes the more uncertain is the reliability. If 
chemists and engineers could put their heads together and produce 
definite conclusions insuring reliability, it would be of the greatest 
advantage. 

The commercial question must also enter largely into the matter, 
and must be considered in each specific case as to whether, in view of 
the higher cost of nickel steel, it may not be of advantage, with bridges 
up to a limited span, to adopt the steel generally used during the past 
series of years. 

No doubt as time goes on and nickel steel becomes better known, 
facilities for its manufacture will be extended and improved, and it 
may replace the older carbon steel. 

One of the uses to which nickel steel may be applied — that of rein- 
forcing concrete — has not been dealt with by the author. It is a use 
to which one can imagine it might be successfully applied, and, if data 
could be obtained as to its freedom or otherwise from corrosion when 
embedded in concrete, it would be most useful, inasmuch as its superior 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 327 

strength and comparative lightness render it suitable for such a Mr. Ross. 
purpose. 

The question as to the value of nickel steel is in a large measure 
a new one awaiting examinatiou and test, and the Engineering Pro- 
fession at large owes Dr. Waddell a debt of gratitude for the able way 
in which he has dealt with the subject. The writer trusts that he will 
continue his investigations. 

L. Dumas, Esq. (by letter). — The author is to be congratulated on Mr. Dumas, 
having undertaken and carried out so well the very important studies 
recorded in this paper. 

The introduction of nickel steel into the construction of bridges is 
one of the most desirable advances, because great opportunities for it 
will be developed, even though it has not yet been adopted for special 
structures of small tonnage. 

The key of the problem is to strike a balance between the great 
advantages of nickel steel judiciously chosen and the increase of cost 
resulting from the introduction of this expensive metal. It seems 
that Dr. "Waddell succeeds in doing this. 

In the following observations, the writer will not discuss the experi- 
ments on bridges or bridge members, as this is not his particular 
sphere; but will speak from the standpoint of the metallurgist. 

The first question presented is the chemical composition of the steel. 
The paper notes, on good grounds, the agreement of technical and 
economical considerations in limiting the nickel contents. They have 
led him to adopt 3^% of nickel for plates and bars. The writer thinks 
there is no possibility of exceeding this percentage; perhaps it is even 
too ambitious to introduce nickel in members which should not be 
annealed. 

In effect, nickel steel should be considered as nickel-carbon-manga- 
nese steel. The phosphorus and sulphur are injurious elements, and 
should always be reduced to a minimum, but the carbon and manganese 
are the constituent elements, active, like the nickel, mainly in in- 
creasing the limit of elasticity. One should sum up the actions of 
these three elements in considering their coefficients, which are far 
from being the same; the carbon is by far the most active, the 
manganese comes next, then the nickel. Account should be taken of the 
effect produced by these additions by considering it as a hardening, 
analogous to that which results from rolling or cold-hammering; with 
this difference, however, that this hardening is not superficial, but is 
produced in the whole mass, as if the molecules of iron were wholly 
compressed by the foreign elements. It is known that the hardening 
increases the elastic limit and the resistance to rupture, and diminishes 
the elongation and the resistance to shock. It is known, also, that it 
improves the quality of extra mild (very soft) steel, but when it is 



328 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Dumas, too intense it becomes injurious. In the same way, an addition of 
nickel is never injurious if less than 2%, but may become an advantage. 

One might be tempted to substitute some carbon or manganese for 
nickel because their properties are similar, but this substitution should 
not be made, partly because, for a like increase in the elastic limit, 
the nickel diminishes to a less extent the elongation and the resistance 
to shock. 

The proportions: Nickel 3.5, manganese 0.70, carbon 0.38, are 
probably the highest which should be adopted for bridges, because the 
dimensions of the parts do not permit of submitting them to be 
annealed, a treatment which becomes necessary in order to diminish 
the interior tensions produced by the introduction of these three 
elements. 

The foregoing are the principal thoughts suggested by this paper. 
It is hoped that the author will pursue his experiments, as they 
appear to have been conducted along the proper lines. 

Mr. Perry. ViCTOR Prittie Perry, Esq. (by letter). — This paper has been read 
with much pleasure, and its subject cannot fail to interest in the 
highest degree all railway and bridge engineers, and to a greater extent, 
perhaps, all practical steel manufacturers. 

It would seem to the writer that the necessity for, and utility of, 
further research are fully demonstrated. Some of the experiments fail 
to agree, others are not quite conclusive, and there may be a couple 
of points which have not been mentioned at all. The facts that have 
been brought to light prove beyond dispute the utility of the subject 
under discussion. 

There appears to be insufficient information collected at present 
with regard to the composition of the alloy. The work done is most 
instructive, as far as it goes, and shows apparently that there will be 
a wide field for the use of nickel steel, but further experiments are 
required in which carbon, manganese, and nickel are varied, and varied 
largely. It is impossible to say what results may not be obtained. 

Dr. Waddell does not say why it was decided to leave the acid 
open-hearth steel out of the question. It would be a pity if any 
recognized method of manufacture of carbon steel should be barred 
for the production of the new alloy, for, if so, it would probably have a 
marked effect on the cost of its production. 

It appears plain that nickel steel is not as ductile as medium carbon 
steel, and the effect of cold on the bending tests is very marked. Were 
any tensile tests made with the alloy at different temperatures? It 
would almost seem that those were required. Nickel steel is stated 
by some authorities to be more ductile than medium-carbon steel. 

It would also be advisable if a series of experiments were made on 
the effect of continued vibration on nickel steel, as the results of 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 329 

such experiments are required in order to fix satisfactorily the working Mr. Perry, 
loads for short spans, bridge floors, cross-girders, etc., if it is ever 
profitable to use the steel in such places. From a practical point of 
view, one would like to know whether there will be any greater diffi- 
culty in straightening members of structures bent in transit in the 
nickel steel than in the carbon steel; and how would it best be done, 
that is to say, cold, or after heating. 

The effect of the acid test would seem to point to .the inadvis- 
ability of using the steel in a dense manufacturing district like 
Lancashire; but the result of the smoke test appears to be curious in 
comparison with the acid test, as it is generally believed that it is the 
sulphurous acid in the smoke that damages structures. 

The result of the torsion test surprised the writer; it would seem 
to show that nickel steel is not in every respect stronger than medium 
carbon steel, but perhaps further experiments with different propor- 
tions of the three principal ingredients will reverse the result now 
obtained. The resiliency tests are also inconclusive, and further ex- 
periments under this head are required. The fact that three of the 
four results favor carbon steel would seem to prove that there was 
something faulty in the methods adopted, and these perhaps could be 
improved and extended. 

It will be necessary to devise some method of fixing definitely and 
reliably the "yield point" and the "elastic limit," and these two terms 
should not be confused. The specified method of determining the 
"elastic limit" would seem, properly speaking, to be more appropi'iately 
styled the method of determining the "yield point." 

Speed in determining the tensile strength is admirable, if it can be 
attained, but it is only a secondary consideration, and the more com- 
plex an alloy is, and consequently more troublesome to manufacture,' 
with correspondingly greater possibilities of error, the greater necessity 
for more painstaking and vigorous tests. The more expensive an alloy 
is, the more time can be spared by the manufacturers, and should be 
demanded by the purchaser, for careful tests, and such tests should 
not be proportionately more expensive for the dear alloy than for the 
cheap one. 

Everything in the experiments seems to point to the necessity for 
specifying very carefully the nature of the tests that will be required, 
and in such specifications the speed of the testing machine and the 
thickness of the specimens will require to be carefully detailed. 

The experiments on columns are among the most interesting fea- 
tures of the paper, but the results are puzzling. From the usual 
formulas for the strength of columns, which have again lately been 
investigated and extended by Dr. Lilley, of Trinity College, Dublin, it 
has always been accepted that colum.ns and struts of the same length 
and similar cross-section, differing only in the material of which they 



330 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Perry, are composed, should have their ultimate strengths differing in the 
same ratio that the compressive strengths of the materials differ. As 
far as our present knovi'ledge (which is not very great) on the strength 
of columns goes, the above law holds, but the tests of columns of 
medium carbon and nickel steel do not seem to bear out this law, 
because the ratio of ultimate strengths of 10-ft. columns is given as 
175, and for 30-ft. columns it is only 147. The ratio of loads pro- 
ducing a permanent set of 0.005 in. varies still more, being 183 for 
the 10-ft. columns and 245 for the long columns. It is to be hoped 
that the experiments already made will be largely extended, and it 
would not be surprising if the ratio given for the failure of the long 
columns should finally be found to be more in accordance with facts. 

No tests except those referring to the "bearing-on-pins" tests 
appear to have been made in connection with the ultimate compressive 
strengths of the nickel and carbon steels; such tests would seem to 
be needed. 

The working stresses proposed for nickel steel, if the elastic limit 
and ultimate strength given are to be accepted, would not be regarded 
as satisfactory in Great Britain. With a practically new alloy, a far 
larger factor of safety should be adopted. American practice, in this 
respect, with regard to medium carbon steel, is so different from that 
in use for many years in the United Kingdom that the views of a 
British engineer will hardly find favor, but in view of some recent 
bridge failures in America, it would seem that the current American 
practice could be modified with advantage. 

The difference between maximum allowable working loads in 
America and the United Kingdom will render it somewhat trouble- 
^some for a British engineer to use the excellent diagrams, which must 
have given the author a great deal of trouble to prepare, showing the 
relative weights and costs of nickel- and carbon-steel bridges. 

The advisability of mixing nickel and carbon steels ultimately in 
bridgework, such as in plate girders, making the booms or part 
of them of nickel and the webs of carbon steel, seems problematical, 
and Dr. Waddell's remarks on the inadvisability of using nickel-steel 
rivets with carbon-steel plates would point in the same direction. To 
the writer it does not seem likely that nickel steel will be used to a 
great extent for bridges of less than 150 ft. span, as there must be a 
certain minimum dead load in comparison with the live load, and, as 
far as the vso-iter's experience goes, he would not care to cut down the 
minimum weights which can be obtained by careful design with carbon 
steel in short-span bridges, using the working stresses regarded as 
suitable in the United Kingdom, 

For bridges of "00 ft., or llioroabout, no doulit the use of nickel 
steel would prove economical, and for bridges of the maximum span it 
will be indispensable 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



331 



In conclusion, the greatest praise and most hearty thanks are due Mr. Perry, 
the author from the Engineering Profession for this very excellent 
paper. The writer feels sure that only a few realize the amount of 
work involved in its preparation. No doubt Dr. Waddell's desire will 
be attained if the paper leads to further experiments, as there is little 
doubt it will. 

W. H. Warren, Esq.* (by letter).- — The writer's experiments on Mr. Warren. 
nickel steels were published in 1898 and 1902.t Recently, some speci- 
mens of this steel have been tested in Charpy's and in Guillcry's impact 
testing machines, using standard notched bars. The material was 
supplied by the firm of Fried. Krupp, of Essen; its composition is 
shown in Table 66. 

TABLE 66. 



Reference 
letter. 


Carbon. 


Silicon. 


Manga- 
nese. 


Phos- 
phorus. 


Sulphur. 


Cop- 
per. 


Nickel. 


A... 


% 
0.37 

o.:m 

0.10 
0.34 
0.566 
0.52 


% 
0.260 
0.200 
0.012 
0.224 
0..338 
0.29 


% 
0.31 
0.29 
0.33 
0.24 
0.49 
0.35 


% 

b'.m 

0.013 
0.016 


% 


6! 084 
0.010 
0.019 


% 

o.im 

0.064 
0.064 


% 
8.05 


B 


5.60 


E 


6.01 


F 


6.17 


G 


25.74 


H. 









The writer was informed in 1900 by Director Herr Uhlenhaut that 
steel plates containing less than 25% of nickel could not be rolled 
sufficiently smooth. The author's exceptionally able and complete 
paper shows that this difficulty has been overcome, and that the cost 
of manufacture has been reduced so much that nickel steel containing 
3.5 and 4.5% of nickel may now be considered seriously as a material 
possessing decided advantages over carbon steel for the construction 
of bridges. The paper demonstrates that, just as structural steel has 
displaced wrought iron, it may be reasonably expected that nickel steel 
will ultimately displace carbon steel in structures. 

Unfortunately, the steels described in Table 66 differ considerably 
from those dealt with by the author, and an exact comparison is not 
possible, but they were tested carefully by the writer with a machine 
shown by calibration to be accurate within 0.04%, and the deformation 
used for the determination of the elastic constants was obtained with 
Marten's mirror extensometer. The results of these tests are generally 
in accord with those quoted by the author, and they are summarized 
in Tables 67 to 70, as they may be considered to be of some interest 
in this discussion. 

* Challis Professor of Engineering, University of Sydney. 

t"Some Physical Properties of Nickel Steel," Proceedings, Royal Society of New 
South Wales, 1898; also, The Australasian Association for the Advancement of Science, 
Hobart Meeting, 1902. 



332 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Warren. TABLE 67.— TENSION TESTS. 



Reference 
letter. 


Tensile 

strength, 

in tons 

per 
square 
inch. 


Limit of 
propor- 
tionality, 
in tons 

per 

square 

inch. 


Ratio of 
limit to 
break. 
Per- 
centage. 


Contrac- 
tion of 
area at 

fracture. 

Percent- 
age. 


Elongation 
measured after 

FRACTURE. 


Coeffi- 
cient of 
elastic- 
ity, in 
tons per 
squai'e 
inch. 


Quality 
factor. 




On 6 in. 


On 3 in. 




E 


82.88 
43.50 
48.00 
.51 .40 

45.3 
45.3 

47.88 


19.8 
33.7 
36.4 

23.8 

30.2 
30.2 
13.1 


59 

78 
76 
47 

67 
67 
27 


72 
61 
63 
43 

58 
56 
70 


1.75 
I On 8" 

) 1.45 

1.40 

I On 6" 

h.30 
\ On 8" 

\ 1.40 

1.45 

I On 6" 

^2.35 


1.08 
On 4" ^ 

1.00 ^ 

1.00 

On .3" ) 

0.80 i 
On 4" 

1.00 f 
0.98 
On 3" 1 

1.36 f 


12 375 
12 500 
12 500 
12 500 

12 7G7 
12 767 
11760 


9.6 


F 


7 8 


F 


8.4 


F 


11.1 


A 


7.2 


A 


8 2 


G 


18.7 







TABLE 



-Compression Tests of Cylinders one inch in 
Diameter. 



Reference 
letter. 


Length, in 
inches. 


Limit of propor- 
tionality, in tons 
per square 
inch. 


Coefficient of 

elasticity, in tons 

per square 

inch. 


Reroarks in regard 
to limit E. 


H 


2 
2 
4 

2 

2 
4 

2 
2 
4 

10 

10 

10 


12.05 
13.30 
15 60 

25.60 
25.60 

22.18 

28.43 
28.43 
20.00 

11.99 

14.50 

13.10 


10 625 
10 825 

10 625 

11785 

11 785 
11785 

11 071 
11071 
11071 

11360 

12 705 
12 570 




H 


do. 


H . 

A 


do. 
do. 


A 


do 


A 


do 


B 


Well defined. 


B 


do 


B 


do 


G 


do 


E 


do 


F 


do 







The coefficient of rigidity was obtained by measuring the twist of 
a bar 15 in. long and 0.723 in. in diameter: 

A = 11 600 000 lb. per sq. in. 
B = 11 200 000 " « " 
H=ll 800 000 " " " 
Table 71 shows the properties of 6.1% nickel steel in which the 
percentage of carbon has been varied, producing three qualities, mild, 
medium, and hard; the complete analysis was not determined. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



333 



TABLE 69.— Torsion Tests. 

Length on which strain was measured = 3 in. for A, B, and if, and 
1.125 in. for G, E, and 1?'. 



Mr. Warren. 



Reference 
letter. 


Diameter, 

in inches. 

d. 


Number of 
turns at 
fracture. 


Twisting 
moment at 
yield point, 
in inch- 
pounds. 


Twisting 

moment at 

fracture, in 

inch-pounds 


Stress at 

yield point, 

in pounds 

per .square 

inch. 


Stress at 
fracture, in 
pounds ]ier 
square inch. 


A 


0.75 

0.75 

0.75 

0.715 

0.715 

0.715 


2i 

2i 


7168 
7 168 
5 380 


7 710 
9 420 

7 168 
10 931 

6 809 

8 960 


86 600 
80 600 
05 100 


93 300 


B 


113 700 


H 


86 600 


G 


152 578 


E 


95 448 


F 


125 591 







TABLE 70.— Shearing Tests. 

Length of piece, 5 in.; grooved at two sections for douhle shear, 
0.16 in. 



Reference 


Area exposed 

to shear, in 

square 

inches. 


Load at 
rupture, 
ill tons. 


Shearing stress reduced to 
single shear. 


letter. 


Tons, per square 
inch. 


Pounds, per square 
inch. 


A 


0.88 
0.88 
0.88 
0.95 
0.96 
0.96 


27.4 

38.25 

25.00 

44.40 

31.92 

42.50 


31.14 
43.46 
28.45 
46.50 
33.20 
44.10 


69 750 


B 


97 500 


H 


63 700 


G 


104 050 


E 


74 312 


F 


98 780 







The results are interesting, more especially as some pieces were 
first subjected to repetitive tests in a rotating machine under an 
extreme fiber stress of 54 085 lb. per sq. in. before subjecting them to 
the tensile tests. 

In regard to the relative resilience of nickel and carbon steels, the 
results obtained by the author are less favorable for nickel than for 
carbon steels, whereas Professor Hatt's tests are decidedly in favor 
of nickel steel. 

The writer obtained the results recorded in Table 72 with a Marten's 
impact testing machine, similar to the one described in Johnson's 
"Materials of Construction." The specimens were subjected to a shock 
producing tensile stress; they were of standard form, L = 11.3 yja, 
the volume of the parallel portion being 0.212 cu. in. It required a 
number of blows to break the piece, and the extensions per blow were 
measured with a cathetometer. 



334 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



Mr. Warren. 



3 H9 



tP 



1*^ o-cti D 2 ^ 
a-, p 



I'm a 



3 ¥ 

' s; r^ — ff 






°^ 2 n 

B O c 3 !^ 

c 2. §■ 5' 0) 



^ r o OS C3 



£,-■' » 
f» 5 S 

CO f^ CD 






o c „ 
i» » B 



M 2 



02&5COCCCCCOCO c.occccosc^coco ^:'^o(o^^^st^ 



3 to-} It 



l« tS bS O <1 o ts 



w m o 6i o o c 



D" c: en to m oo 
-• c o o o o o 



tots i--^^tOMtOJ 



ooooooo 



"5 -3 -^ O -1 -» -J 



tiiiiisiStitr cc o3co*:*:4i*: 



>0 iSri 



_ OOQDOOOiOOOO 
__ ... _ -JO-vIO OS 05 05 en (*». CI» Oi 

oo ooooooo ooooooo 



(E 



. cc or CO >-* Oi CO 
Bi O ~J en ^ iO 

oooooo 



oc;t uicncncn'jfCjTOT »^j^4;t.tu^kj^t^ O'cocccocccoco 

en ^ to 10 ** 4^ rf^ CO 4^ en en en c;» *»- en en t to to eo co co co 

OiCi -J -^ o to :rs Oi .&. ►^ ^ 4^ H^ '-^ K^ »^ X- ^J oc to "^ CO t-^ 

□oQococnoojito -i-3-N»H-en — -j Ji(».>->ootooo 

5' 

^1^ CO CO CO CO CO CO CO to CO CO CO CO CO CO r^ to to to to to to 

t^O *J -*aOQDCC QOOO OOt-^tOOOOO p'OOtOtOtOCO 

lo'tD o bo -5 -J to 03 OS 1-1 '-IO OS OS OS OS * coco'sowcn*^ 

-^*Q -^*j*Q*5*^-'j*^ osos^cscaosos 2osOSOSOS03*3 

50H- O^-OtOJ^tOO 1-' CO O »J 00 -1 -1 2 f* N) CO 00 -Nl k- 

o 

oo CO OS *. OS en >ft. CD •<} o i*^ oo c» oc to p i-" ^^ to 1-4 -j «o 

5' 

00 ooooooo ooooooo iRoooooo 

enos 4^»^4^4-t^i;^4^ co4^*&2'i4^cOht>.'.c^ Rco'coeccococo 

^ to -3 CO <J 00 -3 X 00 cDOcDOQOOtO r? 4^ 4^- ec CO *»- CO 

oo tocci-*ocDOo co*}os4i.eno*3 Pejicn-^^04^ 
o 

B- 

CO ooooooo ooooooo 5'^^'^^'^'^ 

to W ^-4 ►-* 1-1 — i i-i ,-* h-t H- h-i .-• »-i *i-t (-1 'h- CD o *o *o o o o 

CnO "^IJ^-JXODtXOO tOCOtOtO'--t04* • CDCOODOOCOOC 
ejc-^ 4^-.J*kOOOO l-*OCOQOOSOICO rfi.i&COCOO<i 

i-'CD eDOH--J":C;0 COt-tOO^-JtO h-'.Xh-'-'-JOS 

ocn-}0 coenen to to co co co 






CO 

13 H 

o s 



stress, in tons 
per square inch. 

Limit of elas- 
ticity, in tons per 
square inch. 



Ratio of limit 
to break. 



-.7; t> 

O B CD 

."■' S p 



59 

S 2! 



(^ CO *. en ** *^ *». »c^ 1^ ct: en o en OS en c;t 



OS OS -J -3 ~I •<! 



tOtOCOI-'>-'Oi-' 0-^OXtOcb05 • COCDCS'SOI' 
>(».-^ CDOlo'cto'xO ')t»'osO'— 0'0'co ' V-" i-1 bo CO '*. ►- 



oo'^cococococo<o«o isissotocotsi-4 • oscnuicncsen 



ocx„. 0'-'io>b.iticoc3s -aoosi*»-oosos 



OQCfeO-JOS 



Contraction of 
area, percentage. 



oOjjOsosososososcs xcx>ooooxc»oo oo- ocooo 
'oj'it<.'^''io'-l^eni&.V"os 03tobDi*>.'-jIft.'o Ii^* bcntobccn 



ooopooo 

CO if^ CO CO CO CO CO 

.. , m f — 



4^ CO o en 1)^ en OS en>o> 



pp appp c 
4^ rf^ j^ *i j^ t^ [^ 



opoopo 

c *>■ en *cj» 4^ en 
ooto tOi&ac4^ 



H H 
?> H 2 

o * « 

350 

a o 2! 



IK 



Local elonga- 
tions. 



c 1^ COOOOCOOCOO 

(s'iCk OS o CO CO 00:0 






CO to en CO >t^ >u to 
OS OS o 4^ CO 00 



en -.c -J OS CO ~3 

cocoos'o'»o 



osoo cnifecnencnaicn oscn-josososen • oooscnotcncn 
to'io i-''io'-i'os'cni-''i&. to -J MOW CO en • ce c« \o co -j 03 



tOi-'lO 

-i*.o 
OS CO to 

SS8 



General elon- 
gation, per- 
centage. 



Coefficient of 
quality. 



Revolutions 
before testmg. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



335 



TAI5LE 72. — Impact Tension Tests. 
Weight of hammer, 79 lb. 



Mr. Warren. 



Description, or 
reference letter. 


Number 

of 
blows. 


Total extension, 

up to the last 

blow before 

fracture, 

in inches. 


Total extension, 

measured after 

fracture, 

in inches. 


Mean extension 
per blow, 
in inches. 


Total extension, 

percentage on 

3 in. after 

fracture. 


Work done in 
causing rupture, 
in foot-pounds. 


S5 a y 

•9a.a 

C O 3 


H 


C 

7 

8 
6 

6 

7 
17 
f7 

t8 


0.756 
0.681 
0.796 
0.645 
0.640 
0.680 
0.592 
0.664 
0.916 
0.872 
0.656 
0.692 




0.151 
0.113 
0.1.33 
0.092 
0.128 
0.126 
0.118 
0.110 
0.153 
0.145 
0.094 
0.099 




787 

918 

918 

1040 

787 

787 

787 

918 

918 

911 

1049 

1049 


3 712 


H 


0.90 
0.90 
0.80 
0.70 
0.90 
0.70 
0.75 
1.00 
1.00 
0.70 


30.0 
30.0 
26.6 
23.3 
30.0 
23.3 
25.0 
33.3 
33.3 
23.0 


4 330 


H 


4 330 


A 


4 938 


A 


3 712 


A 


3 712 


B 


3 712 


B 


4 3.30 


B 


4 330 


Axle steel 


4.330 


Tire steel 


4 938 
4 938 





The following experiments were made, with a hammer weighing 
122.5 lb., on the nickel steel recorded in Table 72 with one blow, and 
the mean specific impacts in foot-pounds per cubic inch for the mild, 
medium, and hard varieties were 1890, 1690, and 1080, respectively. 

The same machine was used to obtain the results recorded in 
Table 73, in which the specimens consisted of notched bars, and the 
weight of the hammer was 40 lb. The notch consisted of a hole formed 
with a twist drill 4 mm. in diameter in the specimens 20 by 20 mm. 
in cross-section, and 8 mm. in diameter in the specimens 30 by 30 mm. ; 
the cross-sectional areas at the notches are 200 and 450 sq. mm., re- 
spectively. The specimens were supported on a span of 120 mm., and 
were subjected to a number of blows varying from 0.15 to 1.00 m. 

In impact tests it is necessary to keep all the conditions constant, 
which is accomplished by preparing all the specimens of the same 
form and dimensions, and testing them in the same machine under pre- 
cisely similar conditions, such as weight of hammer, height of drop, 
and equal intervals of time between each drop; the specific impacts 
will then represent the relative resistances to shock of the materials 
tested. 

The Charpy and Guillery impact machines are now much used in 
the writer's laboratory on specimens having a length of 60 mm., a 
cross-section of 10 by 10 mm., and a square notch 2 by 2 mm. The 
energy absorbed by the test piece in a single blow is determined 
accurately, and also the angle of rupture. 

The results obtained with the steels denoted by A, F, and H, in 
Table 66, are as follows: 

H gave 5.2 kg. and 5.5° angle of rupture. 

A " 16.0 " " 26.5° " " " 

F " 21.2 " " 29.1° " " " 



336 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



Mr. Warren. 






!*i .'*! ^ ti. S >" 4: .•*) tt. H t> £q ^"^ tt. |k 



O l-i O — O k-* O I-' O _>-' 

en o CT Q ui o cjt o or o 



ocoooooooo 



coo^^50^coaD•^oc^>*^oocccnoi■o^f*.&^^lcnm 



ooooocooooooooooooooo 
K) to en o: M Vo Qo 'o> M OS '-- *. ►— -J ai ►-' It' ^^ o; S 2 



ooooooooooooooooooooo 
1 o o o o o o o o o ► 



ooooooooo 



8£ 



,Jib.f0GCO5O5 

0Da0Ol4^K0Cp«>.OOO 



CO CO -3 -} It' M ■-> to 05 05 i-i tSi-* -i iS i-i "^ 

SCO 00 00 ci» ?c cc I— a» cjf *> or CJ» «3 -^1 -^> cf» -^j o: to o 
•J J-s»504-.0'.CO;OSOC-^:CCJ»000'iCOt5CiO: 

OOOo"cnoOC0t^'w*>it'OWO<J4i.*?CnXQD 



a ^S? 
p 1 2 

?=PO 



o 



. P CP «! 
! o &ai 
P (C _ 



C T I— <-^P 



™ s o ''S 

en ? M * 



•c!; 



o I— tpS 



■OOrtj 1^ 3* 

O "_. n^ "! 

6 D P !^ 

«2 p o 

• O ^ 3 



o 
o 

H 

o 

w 

K 

o 

td 
> 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 337 

Unfortunately, these results on shock resistance are not strictly Mr. Warren, 
comparable, but, taken as a whole, they show that nickel steel possesses 
' considerable shock resistance, at least equal to that of carbon steel. 

In regard to the so-called brittle zone between 5 and 12% of nickel, 
this does not appear to be due to the nickel, as Eudeloff showed in 
1896 that the ultimate strength, elastic limit, and yield point increase 
with the percentage of nickel up to 10%, after which they decrease. 
There was no suggestion of brittleness in the writer's experiments of 
6% nickel steel, so that the brittleness found by other experimenters 
must have been due to other causes, such as the greater proportion of 
carbon and manganese used. 

In regard to resistance to corrosion, the writer used a 1% solution 
of sulphuric acid, and boiled the specimens for four days. The results 
obtained showed that nickel steels containing 3 and 6% of nickel were 
at least as good as carbon steel. The 25% nickel steel was perfectly 
bright after the test. 

In reference to the composition of the rolled steel given by the 
aiTthor in Table 4, the writer believes that a suitable material for 
bridge work might reasonably be expected from this specification, and 
the tensile strength and elastic limit should also be easily obtained. 

In regard to the intensities of working stresses in bridges, the 
writer considers that the practice of using a suitable impact formula 
to express the effect of the live load, and the reduction to a total 
equivalent dead load is thoroughly sound, and most convenient in 
designing. The intensities given by the author are fully justified by 
the various tests of nickel steel. 

The most satisfactory method of finding the safe working stress in 
nickel steel is to multiply the known safe working stress in carbon 
steel by the ratio of the limits of proportionality of the nickel and 
carbon steels. 

The compression formulas also appear to be consistent, having 
regard to existing knowledge on the strength of long columns, although 
further experiments appear to be desirable. 

The writer has read the paper with great interest, and is impressed 
with the author's comprehensive and thorough treatment of the whole 
subject. The diagrams giving the cost of different types of bridges 
suggest that nickel steel will probably soon displace carbon steel, in 
whole or in part, for the construction of railway, and for long-span, 
bridges. 

William E. Webster, M. Am. Soc. C. E. (by letter).— The only Mr. Webster 
matter in this paper the writer cares to discuss at this time is the 
specification for nickel-steel eye-bars. 

The author's requirement for bending tests on unannealed speci- 
mens, of 90° around a pin having a diameter equal to three times the 



338 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



Mr. Webster. 



















-• 3 






















c p 






10 


ts 


ts 


J_l 


t_l 


,_t 


J_^ 










if: 


J^ 




^ 


^ 


s: 
























w p 


















































^ 


in 


en 


en 




OS 


05 


OS 


s-^ 




•<i 


^ 


^ 


C3D 


o 


N. 


it. 


s 








o 


o 


o 


O 


^ 


O 


o 


Q 
















o 


o 








° 


^ 


° 


° 


° 


o 


o 


° 


• o 


'^ 








































I 




S° 


to 


o 


to 


o 


^ 


^ 


^ 


C 


o 

3 




tn 


(X 






-V* 


o 


05 


en 


^ 




OI 








CJI 




en 










o 






o 


o 




o 




05 


C6 




o 


° 


° 


o 


o 


° 


o 


° 


f 












































»_i 


H-i 


,_i 


,_! 


(__. 


y_. 


j_i 


,_^ 


;3 






o 


w 


05 


s 


g 


^ 


g 


g 








CJl 


or 


Of 


CT 


en 


CJI 


CJt 




OR 










o 


O 


o 




o 


ct 






"-■ 


° 


o 


O 


o 


o 


o 


o 


P" 




a; 










































K- 










































M 


Cl 


OI 


en 


Crt 


en 


05 


05 


05 


-W 




o • 

05 

•a 


-J 


-J 


-J 


-^ 


00 


o 


iO 


M 


II 




o 


o 


o 


o 


^ 


(3 


o 


o 




s 


CD 

o 


o 


o 
o 


s 


8 


g 




■ o' 












































eq 




















O 


o 




















to 


to 


to 


o 


£ 


o 


^ 


^ 


a 


3 




en 


U1 


00 




-J 


o 


to 


fT 


Ci 


te 


o 


2! 


ai 


in 


en 


en 


en 


o 


3 


s 




8 


o 


s 


o 
o 


8 


g 


g 


s 


d 






















(T) 




















fD 


.-) 










































^^ 






















^ 






















(B 












h-A 










13 
















to 


^ 


to 


•s 






o 


O 


C6 


03 


to 


to 


CI 


^ 






o 


OI 


CJ» 


en 


en 


en 


en 


1— , 


s 








o 


o 


o 


o 


o 


o 










° 


o 


o 


o 


o 


o 


o 


"^ 


























;? 


















1— 1 

a 


», o? 
p """o 




o 


5 


s 


^ 


05 


oJ 


>(>. 


ts 


GO 






















B 






















• ' OK! 






















(t 




8 


B 


s 


to 


s 


^ 


to 


8 


CO 


W 
d 




















p. 




lb- 


OI 


Oi 


en 


en 


en 


en 




c-S 






CO 


o 


^^ 


eo 


en 


02 


-1 




5"p 






or 


CO 


*o 


OS 


Q 


CO 


-J 




la 


^ 






^ 




K) 




^ 


en 






o 


o 


o 


en 


o 


o> 


o 




• n 


c 
do 


























































a-a 


N 




















■^ t? 


ei 




to 


to 


to 


to 


g 


o 


o 




(6 2" 




o 


to 


OI 


~i 


tl 


Ol 




5i 


d 




g 


o 


g 


en 
o 


g 


o 


o 




IS 






o 




o 


° 


° 


o 




FS 






















Percent 

of elo 

gation 

10 in 


o 


o 


c 


o 


o 


o 


o 
























• -.l?«o 






















d (W 






















a> 







DISCUSSION ON NICKEL STEEL FOR BRIDGES 339 

thickness of the bar, is not rigid enough to insure the proper quality Mr. Webster, 
of steel being used for eye-bars, or to insure care in rolling. 

He starts with a leeway in ultimate strength of from 115 000 to 
130 000 lb., and treats the whole matter as though one heat of steel 
should be rolled into bars of any thickness that might be required. 
Allowances are made to cover this, both in the requirements of the 
specimen and of full-sized eye-bar tests. The permissible variations 
are worked out in Table Y4. 

The differences in ultimate strength in specimen tests are increased 
to 33 000 lb., or more than doubled, the low limit for a 2i-in. bar 
being 95 500 lb., and the high limit for a 1-in. bar being 128 500 lb. 
This is too great ; 15 000 lb. is enough leeway in any specification for 
nickel-steel eye-bars. 

These specifications really allow the use of two grades of steel, 
owing to the wide permissible variations in ultimate strengths. The 
author also calls for much higher ultimate strengths and elastic limits 
than he obtained in any of the eye-bars he tested; and he gives what 
he thinks would be desirable. He refers at length to the results of 
tests of eye-bars that met the requirements of the specifications for the 
Blackwell's Island Bridge; had he based his conclusions on the results 
of tests of eye-bars that did not meet the requirements of that specifica- 
tion, he would not have specified any material up to 128 000 lb. in 
tensile strength. 

Particular attention is called to this, as some may be misled 
by his specification, and have trouble with this higher tensile-strength 
steel. 

William H. Breithaupt, M. Am. Soc. C. E. (by letter). — This Mr. Breit- 
paper, containing the results of investigations made at considerable ^^"^ ' 
cost, and extending over several years, is a valuable addition to engi- 
neering knowledge. It is a description, tabulation, and graphical ex- 
position, in Dr. Waddell's well-known exhaustive manner, of properties 
of nickel steel, with a view to demonstrate its economy as a material 
for bridges, and to find the alloy best adapted as such material. With 
due appreciation of the facts brought out, however, it must be held 
that the argument for the general use of this material for bridges 
remains open to various objections. 

The tests show that the elastic limit of the alloy of nickel steel used 
is about one and three-quarters times that of medium-carbon steel, 
that the ratio of respective ultimate strengths is slightly less in favor 
of nickel steel, and that the modulus of elasticity is about the same 
for the two materials. Some of the bending and shopwork tests were 
favorable to nickel steel and some inconclusive. The corrosion 
of nickel steel is probably less than that of carbon steel ; the tests on 
this were conflicting, and do not give much definite evidence. The 
alloy of nickel steel used is less ductile than the ordinary low-carbon 



340 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Breit- steel, and therefore does not stand drifting as well. Impact tests 
showed better for carbon steel than for the nickel steel used. Shop- 
work on nickel steel will require heavier tools throughout, as shown 
by the punching, shearing, reaming, and chipping tests. With an 
alloy containing about 4^% of nickel there is danger of brittleness, but 
this is not well defined. 

The testing of nickel steel requires to be done very carefully. 
T?esults differ materially, depending on the size of the rolled sections 
and on the location of the test piece, whether from the edge, body, end, 
etc., of the rolled section, the difference thus revealed being much 
more marked than in carbon steel. 

The superiority of nickel steel for eye-bars may be considered as 
well established. For the same work, nickel-steel eye-bars can be more 
than 40% lighter than if of carbon steel. With the greater uncertainty 
of testing, proper safe-guarding, however, would require the imposing 
of the old-time specification of the initial testing of each individual 
finished bar, and this should be to a load well above the elastic limit 
of carbon steel. Such a requirement was not uncommon at the time 
when steel eye-bars superseded wrought-iron bars, and when it was 
difficult to get uniform steel. It would not be practicable for com- 
pression members. 

The compression tests of struts, Table 27, show the superiority of 
nickel steel over carbon steel as considerably more for long struts than 
for short ones, within the limit of set, i. e., within the elastic limit of 
the material; while, at the point of failure, the results are relatively 
reversed. The values are as 2.44 nickel to 1 carbon for the 30-ft. 
struts, and as 1.83 nickel to 1 carbon for the 10-ft. struts, within the 
elastic limit, and respectively at 1.76 to 1 for short, and 1.47 to 1 for 
long, at the point of failure. With the practical equality of E for both 
materials, the superiority of nickel steel for use in compression will 
stand more investigation by tests of full-sized bridge members before 
it can be considered that definite values have been obtained for the 
proportioning of sections. It is a grave question whether our apparent 
over-confidence in steel for large compression members is not due to 
insufficient consideration of the fact of its relative lack of superiority 
in stiffness. 

For smaller compression members, there appears to be little or no 
advantage in the use of nickel steel, and the same may bo said as to 
lloor-beams and stringers — where impact, among other things, comes 
into play — and for girders generally. Whether the economy in weight 
of long compression members would figure much more than 20% over 
carbon steel, must be held to remain in doubt. Such economy, how- 
ever, is material and will, with the much greater saving in eye-bars, and 
with greater facility in its production, give a definite field for the 
use of nickel steel in long-span bridges, and particularly in bridges of 
exceptional span. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 341 

The author is entitled to the gratitude of the Engineering Pro- Mr. Breit- 
fession for his timely investigations and for so well putting them on ^^^ ' 
record. 

E. A. Stone, Esq, (by letter). — On reading Dr. Waddell's paper Mr. stone. 
one cannot but be struck with the heavy task which he laid out for 
himself in the beginning of his investigations, and the thorough 
manner in which he afterward carried it out. 

At the present time, when our long-span bridges appear to have 
reached their maximum length, a thorough investigation into the possi- 
bilities of "Nickel Steel for Bridges" would seem to be most opportune, 
and a material which, by virtue of its high elastic limit and ultimate 
strength, reduces the d^ead load of such bridges, and enables us to 
increase the maximum allowable span by 500 ft., is one which should 
commend itself to all bridge engineers. 

As the paper would seem to prove very conclusively the superiority 
of nickel steel over carbon steel, from the standpoint of the economics 
of engineering, the question now to be answered is: Why not use it? 
The main difficulty, however, would seem to be the sensitiveness of the 
public generally to do anything in the nature of an experiment, 
together with the change of plant required in the bridge shops, the 
present rnachines for carbon steel not being sufficiently heavy ; a change 
which would take place immediately, no doubt, on the demand for 
such a product being made. 

Table 5, comparing probable shop costs, is interesting. As the 
items for drawing-room work, template-shop work, and laying-out 
work show an increase, the writer supposes the table refers to "mixed 
steel" bridges. It would seem that the fabrication of mixed-steel 
bridges might not possibly be regarded altogether with favor by manu- 
facturers on account of the necessity of using two different grades of 
material in one bridge, which would increase the likelihood of error 
in both shop and office over those where the entire structure is made 
from one class of material only. 

An element tending to keep up the cost of nickel-steel bridges is 
referred to on page 242, where the fact is stated that the high tensile 
strength of this material, by diminishing the size of the sections, would 
diminish the bridge shop tonnage, while the operating expenses would 
remain the same. This fact would undoubtedly be a factor in the 
new shop costs to be determined. 

The next logical step would seem to be for some one to build a few 
nickel-steel and mixed-steel bridges, in order to see how the observed 
costs actually compare with those we now have for carbon steel. The 
exliaustive set of physical tests carried nut would seem to be conclusive 
as to the suitability of the material, so that the only "non-proven" 
qualification would be that of actual cost. It is to be hoped that such 
bridges will shortly be built and that Dr. Waddell's cost diagrams will 
be fully substantiated. 



342 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Codron. C. CoDRON, EsQ. (by letter). — The writer has read this excellent 
paper with much interest. 

The results of the author's numerous tests, the observations which 
he has made, and the conclusions resulting therefrom — in fact, the 
whole of this remarkable work — comes at the proper time, and is an 
important contribution to the general study of nickel steel for bridges 
as compared with the steel ordinarily used in metallic structures. 

This paper seems to be very complete, as well as most instructive, 
and deserves the serious attention of constructors. Little remains 
to be brought out with reference to those properties of the steel which, 
within the percentages of nickel and other foreign matters considered, 
best qualify the metal for the uses intended. 

It is desired, however, to call attention to the fact that it would be 
of great advantage to be in possession of additional data relative to the 
manufacture of nickel steel, in order that constructors may gain con- 
fidence as to its uniformity, an essential condition for its use with 
the guaranty of unquestionable success. It would be desirable, there- 
fore, to have data as to the exact choice of the raw materials and, in 
great detail, their mode of treatment: on the casting of the ingots, 
their form, weight, and whether or not they are compressed, what 
defects they show, and the manipulations they must undergo before 
rolling. 

The best methods for transforming the ingots into plates or sections 
are also most interesting studies. It would be of great interest to 
possess detailed information as to the care to be taken in rolling 
nickel steel as compared with ordinary steel : as to the temperatures 
of rolling; how to conduct the drawing out of the metal, whether by 
hammer, press, or rolling mill, with the extreme temperatures during 
these manipulations, and the defects which are developed when the 
methods are not of the best. All these points in the study of nickel 
steel are most interesting, not only with respect to bridges, but also 
all other possible ways in which it may be used, the same methods 
of mill manipulation being required in all cases. 

In France, there are no bridges of nickel steel; engineers have 
been satisfied, heretofore, in making very modest tests on beams with 
reduced dimensions, because it has been difiicult to roll large-enough 
flanges. It would seem that the difficulty of rolling the shapes required 
in bridge construction would lead to the adoption of a steel either free 
from nickel or containing a smaller percentage than may be con- 
tained in steel for plates and other simple sections. 

The author's numerous tests confirm the general facts which have 
been indicated by more modest experiments, particularly as to the 
coefficient of elasticity, ultimate strength, ductility, resilience, and 
resistance to corrosion. It cannot be doubted that for long-span 
bridges nickel steel will assert itself, not merely because it will permit 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 343 

their extension, but chiefly because it deteriorates less than ordinary Mr. Codron. 

steel under the influence of the environment in which it is placed. 

This is a vital quality, which will lead to the use of nickel steel with 

as high a percentage of nickel as possible for bridges of small span 

also, and, in fact, for all structures exposed to corrosion. Since the 

corrosion becomes less as the percentage of nickel increases, the 

main problem will really be to manufacture steel containing a high 

percentage of nickel with as much ease as ordinary steel, and, at the 

same time, give it, not merely a higher ultimate strength, but also a 

high ductility. 

Of the numerous tests which the author has considered, and has 
carried out with great perfection, the writer only wishes to offer some 
remarks on those of flexure or bending, which are in every case 
essential tests for a knowledge of the ductility, one of the most im- 
portant properties of the metal, when combined with sufficient tenacity. 
It is strange that in bending tests it is generally considered sufficient, 
on the one hand, to specify the angle of bend, without even indicating 
the radius of curvature — which shows absolutely nothing — or, again, 
to fix the angle, and also the radius, in terms of the metal thickness — 
which is better, but yet insufficient for rapid comparison with tests 
involving different thicknesses and therefore different radii of curva- 
ture; and this in spite of the fact that it is so simple to obtain the 
important characteristic result of this test, namely, the elongation of 
the extreme fibers at the time when cracks appear. 

This test is most valuable. It should always be made with great 
care, as the author has indicated by placing it among the conditions 
prescribed by the specifications. 

When the bending is made around a mandrel having a diameter 
equal to n times the thickness of the specimen, while its two parts form 
an angle, a, before cracks appear on the extreme face, it is apparent 
that the process described will be sufficient for the purposes of the 
test, but it is useless to specify the angle, a, as the curvature of the 
mandrel itself sets the limit to the bending of the test piece. It may 
be added, however, that it would be very useful to determine the 
amount of elongation, and to note it in the records of the test. All 
that need be done, in order to obtain this value, is to draw on the 
extreme face of the specimen a few transverse lines, at equal intervals 
of say 10 mm. (0.4 in.), measure the distance between these lines at 
the end of the test, and then compute the percentage of elongation. 

It is advantageous to vise first a mandrel having a radius large 
enough to make the angle between the two legs of the test piece less 
than 90°, and then gradually effect the flattening of the specimen, 
preferably by machine. This method permits observation as to the 
behavior of the extreme fibers. The operation is stopped as soon as the 
cracks appear. 



344 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Codron. If care is taken to draw transverse reference lines every 5 or 10 mm., 
it is easily possible to note the elongation of the extreme fibers, and to 
commute its percentage, which may then be compared with the per- 
centage given in a tension test to rupture. 

There is ordinarily little variation between these two values. In 
this manner the bending test may be made with all possible care; 
one may utilize an apparatus for testing beams simply supported, 
such as is easy to arrange for the purposes of this test. 

Where the bending of the specimen is continued until its parts are 
brought flat against each other, it may happen that the bending 
obtained does not give the extreme elongation which corresponds to the 
formation of cracks. In such a case, if it is desired to obtain this 
extreme value, the specimen may be notched on both its parallel 
opposite faces; the notches may be a few millimeters deep, and a few 
centimeters long, and made so as to localize the strain, and obtain a 
greater elongation of the extreme fibers. This procedure has been 
adopted successfully by M. Breuil, Chief of the Testing of Metal in 
the Conservatoire des Arts et Metiers de Paris. 

Mills and testing laboratories should possess a special machine for 
bending test pieces, and for enveloping them around the mandrel until 
they fail, without paying much attention to the angle of bend. 

The author's impact tests, although giving somewhat contradictory 
results, show the difiiculty of obtaining uniformity of fabrication; 
altogether, they indicate important facts which must be taken into 
account in the consideration of nickel steel and its uses. 

The author has been careful to call attention to this point, and 
to those results of the tests which are at variance with current ideas 
on the matter. 

It would therefore be well to continue the impact tests on an 
extended scale, with specimens containing varying proportions of 
nickel, and with fabricated members of the same composition. 

But the important point has been to establish the fact that the 
nickel steel considered possesses a ductility entirely sufficient for its 
use in bridges, and that it ofi^ers complete assurance against brittleness. 

The author's remarks about rivets are very judicious; the writer 
agrees with him, particularly as regards their increase in diameter 
for ensviring greater strength in the riveted connection. 

It is certain that, as a rule, constructors will not adopt nickel steel 
unless the metal is demanded by their clients. 

Engineers will not hastily relinquish ordinary steel, the nature of 
which is so well understood and presents such great uniformity of 
production, for another metal such as nickel steel containing even a 
small percentage of nickel, the manufacture and fabrication of which 
are still subject to many uncertainties. 

The future belongs to the metal which combines high resistance 
with sufficient ductility. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 345 

Mr. Watldell is to be congratulated on having undertaken these Mr. Codron. 
new tests for the purpose of giving more complete information in 
reference to the advantages of nickel steel in bridge construction. 

It is hoped that these most valuable tests may serve to advance 
considerably the study of this problem. 

W. W. K. Sparrow, Assoc. M. Am. Soc. C. E. (by letter). — The jir. Sparrow, 
bending tests to which the material was subjected were severe, and 
4)rove conclusively that, as regards ductility, nickel steel is satisfactory 
for bridge building. The British Standard Specification for structural 
steel for bridges, which is for a 60 000 to 70 000-lb. steel, is that the 
piece must withstand, without fracture, being bent over until the sides 
are parallel, and the internal radius is not greater than one and one-half 
times the thickness of the test piece. The Munich Conference recom- 
mended that the pieces be bent over a rounded edge of 1 in. diameter 
for all thicknesses. 

The writer would certainly expect to find the resilience of nickel 
steel considerably greater than that of carbon steel, and is at a loss 
to account for the contradictory results of the author's tests and those 
of Professor Hatt, unless they can be attributed to the manner in which 
the tests were made. 

The paper does not contain sufficient data to enable one to arrive 
at a very clear conception of the modus operandi of the author's tests. 
His apparatus was of the impact-hammer type, the weight of the 
hammer multiplied by the height of drop plus the deflection of the 
test piece being the work done under each blow, and the ultimate 
resilience of the piece the total work done in producing rupture. The 
rise of the elastic limit after the release of a stress in excess of it is 
well established, and, in cases of repeated impact, stiffening of the 
material takes place; consequently, a considerable amount of the work 
done by each blow is spent in heat due to the change in the molecular 
structure of the metal, and therefore the total work expended is not 
a true measure of the ultimate resilience. In the fourth series of tests 
the fall would appear to be too great, giving a high velocity of impact, 
resulting in the work, to a certain extent, being spent locally, since 
time is required to transmit the internal stress. It would also be 
interesting to know whether there were any moans of measuring 
the rebound of the hammer, as such rebound would naturally have to 
be deducted from the total fall of the hammer in order to give an 
accurate measure of the work done. It is a pity that an autographic 
diagram of the stress-strain curves of both materials was not printed 
with the paper, from which the comparative amounts of work done in 
rupturing the test bars could be accurately computed, but an approxi- 
mation can be made from other observations. Neglecting the small elastic 
work, the stress-strain curve may be considered to be a parabola from 
the yield point to the point of maximum load. Then if A be the total 
elongation measured after rupture, in inches per inch, and Sy and Sm 



346 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

VI r. Sparrow, the Stresses in pounds per square inch at the yield point and point of 
maximum load, the total work per cubic inch up to rupture is nearly 

U^S^X + j (S,„ - S^) X = j(S^ + 2 SJ. 

This equation was first given by Dr. Kennedy.* According to 
Table 28, Appendix A, Sy and S^ for nickel steel are 56 000 and 
104 600 lb., respectively; and for carbon steel, 22 500 and 58100 lb.; 
the elongation, in inches per inch, being 0.173 and 0.308 for nickel and 
carbon steels, respectively. Solving the above equation for 

nickel steel, C7 = 15 293 in-lb. 
and for 

carbon steel, U = U 188 " 
In other words, for 1 cu. in. of material, the ultimate resilience of 
nickel steel is 15 293 in-lb., and of carbon steel, 14 188 in-lb., nickel 
steel being barely 8% higher than carbon steel. 

Of course the writer is aware that the result obtained by a static 
test may differ considerably from a rapid one, and that the resilience 
thus determined is not a measure of the capacity of the material to 
withstand a shock or sudden blow; but the work done in rupturing a bar, 
depending as it does on both the ductility and the strength of the 
material, is a valuable criterion of its suitability for structural purposes. 

Solely on account of the refinement of the testing machine at 
Professor Hatt's disposal, the writer would prefer his tests, and in his 
opinion it is a decided advantage to test by a single blow — as was the 
case in the impact-tension tests conducted by Professor Hatt — than 
by several repeated ones, as the rupture of the bar is caused by a 
known quantity of work, and the indeterminate phenomena, due to 
stiffening and loss of work by heat, are avoided to a great extent. 

It is to be hoped that further tests will be made at an early date, 
and the question of the resilience of the two materials finally settled, 
for the author's impact tests, while very satisfactory as proving the 
great abuse to which nickel may be subjected, can hardly be classed as 
such as a measure of the resilience, and, in the interests of the Profes- 
sion, it is essential that a further series of tests be made. It is to be 
sincerely hoped that some one as self-sacrificing as the author will be 
found ready to perform them. 

In his tests on rivets Dr. Waddell learned that nickel-steel rivets 
should be used in nickel-steel plates and carbon-steel rivets in carbon- 
steel plates, owing to nickel steel being harder than carbon steel. 
This being so, it would be interesting to know how he intends to make 
the connections in his proposed composite structures of nickel and 
carbon steel, for if a carbon-steel member is to be connected to a nickel- 
steel member, the connecting plate must be either of carbon steel or 
* Minutes of Proceedings, Inst. C. E., Vol. LXXVI. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 347 

nickel steel, in which case there will be nickel-steel rivets in a carbon- Mr. sparrow, 
steel plate or vice versa. 

In venturing any further remarks, it would be to note the great 
advantages of nickel steel over carbon steel for long-span bridges, 
whereby it is possible to increase the maximum limiting span of carbon- 
steel bridges by nearly 12%, as shown by the tables, without increasing 
the cost per foot run; and also that the maximum limiting span of 
nickel-steel bridges is 28% greater than that of carbon-steel bridges, 
with an increased cost of 37% per foot run. The advantages in favor 
of the new material appear to be pretty conclusive, with the exception 
that many points have yet to be decided, upon which the author invites 
co-operation, such as the proper treatment of nickel steel to give the 
best results for eye-bars, the best composition for rivet steel, and the 
behavior of the metal under impact. 

In conclusion, those features of the paper which impress the writer 
most are: (1) there is nothing left to chance; (2) there is an open 
mind all through, and no bias; (3) there are no hastily drawn conclu- 
sions ; and last, but by no means least, the whole is done in the interests 
of the Profession, a characteristic far too rare nowadays. 

B. J. Lambert, Assoc. M. Am. Soc. C. E. (by letter). — Assuming Mr. Lambert, 
that the weight and cost curves as given by Dr. Waddell are correct, 
there is then furnished good evidence that for all except very light 
'bridges, the nickel-steel structure is more economical than the carbon 
steel, and especially so when the cost of erection and freight rates 
are high. 

From inquiry among bridge manufacturers the writer has found 
that highway-bridge work comprises probably from 30 to 50% of the 
bridge output of the Central States. Until the possibilities of carbon 
steel are more fully utilized than at present, the use of nickel steel in 
structures of this class would hardly be desirable. It might serve to 
give the scalper another opportunity to boost his already extravagant 
prices, and to thin up sections that are already too sylph-like. 

In smaller bridges there are generally members, the cross-section 
of which Is determined, not from the computed stress, but from the 
allowable radius of gyration, minimum area, or minimum thickness. 
The use of nickel steel in these members would now cause unnecessary 
expense. It might easily be argued then that for these structures a 
Combination bridge composed of nickel-steel and carbon-steel members 
would be desirable. Theoretically, this would do nicely, but in practice 
it is easy to see what troubles would be imposed upon the manu- 
facturer. The exterior appearance of the rolled section would not 
identify its composition, and in the process of fabrication there might 
easily occur an exchange which later would prove disastrous. 

This brings up the point which, it seems to the writer, will have 
considerable weight with the rolling mills and bridge companies. 



348 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Lambert, namely, the difficulty of handling the two steels without danger of 
confusion. This may seem more fanciful than real, but even Dr. 
Waddell says, in reference to assembling a certain strut, "Extreme 
precautions were taken at every step to keep the two steels separate." 

To a certain extent this same confusion may have existed when the 
change was made from wrought iron to steel. It will be recalled that 
the late Colonel Koebling's original plans for the Brooklyn Bridge 
called for wrought-iron cables; but progress in steel-wire manufacture 
caused him to change to steel; and it is due to this change alone that 
the Brooklyn Bridge is now able to accommodate the increased loadings 
imposed upon it. 

Nickel steel has been and is in a somewhat analogous position. Its 
superiority over carbon steel in suspension, cantilever, and long-span 
bridges is apparent, and, with such an enthusiast as Dr. Waddell 
behind it, it is thought that the objections to its use in the more 
common structures will finally be removed. 

There are many points of extreme interest to a bridge engineer in 
Dr. Waddell's paper, and one cannot but be impressed by the energy 
and zeal which produces and then makes public this great amount of 
valuable information. 

Mr. Marriott. WiLLiAM Marriott, Esq.* (by letter) . — ^The information given in 
this paper is certainly very interesting. A steel half as strong again 
as carbon steel makes one long for a chance to try it. We have some 
knowledge of the behavior of nickel rails, nickel steel in high-class 
machine work, etc.; but, so far as the writer is aware, its behavior in 
actual bridge work is not known. It seems that we require to know 
something of its action under "fatigue," and whether it tends after 
years to become crystalline and fracture. 

There is also in Great Britain the great question of corrosion and 
consequent up-keep. It is only now being generally admitted that 
in this respect steel is not as durable as iron, and needs greater care 
in maintenance. 

Steel segregates badly, and the pitting is generally attributed to 
this segregation. If nickel steel will corrode as badly as ordinary mild 
steel, and as rapidly, the pitting on the much lighter scantlings will 
constitute a serious danger. 

As far as the writer is aware, there are at present no Board of Trade 
regulations which would allow of this steel being used in Great Britain 
with economy. 

Mr.Rohwer. IIenry Eoiiwer, M. Am. Soc. C. E. (by letter). — Dr. Waddell's 
paper cannot fail to be studied with great interest by all engineers, 
especially by those who are called upon to design bridges, and they 
will find it, not only of interest, but of great value, should the con- 
clusions reached by the author prove to be true. 

*M. Inst. c. E. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 349 

The author, by liis diagTanis and numerous tests, shows the most Mr.Rohwer, 
efiFective combination, as far as he has discovered; he goes further, 
and, by taking into consideration the price at which .nickel is being 
produced, shows the cost of nickel steel for various styles and lengths 
of bridges. lie has thus contributed to engineering literature one 
of the most important treatises on the application of nickel for techni- 
cal purposes, the extent of such application depending greatly upon 
the quantity of nickel mined and the price at which nickel steel can 
be manufactured. 

Up to the present time, very little has been made known regarding 
nickel steel and its behavior when subjected to strains such as exist 
in bridges, though the idea of its use in bridges is not altogether new. 

When the construction of the bridge across the Rhine, at Cologne, 
Germany, was under consideration, the use of nickel steel was not only 
suggested, but extensive tests were made under the auspices of the 
machine works of Augsburg-Nuremberg and at the factory of Krupp, 
who was to furnish the material. Whether the results were too 
meager, or of insufficient weight to warrant the introduction of nickel 
steel, the writer does not know; suffice it to say that it was not used;, 
and, as far as the writer knows, the results of the tests were never 
published. 

German engineers seem to consider it safe to use nickel steel with 
from 2 to 8% of nickel, for bridges, for it is claimed that such 
material has shown a tenacity of 65 kg. per sq. mm., a ductile limit 
of 45 kg. per sq. mm., and a flexibility of 18 per cent. Compare 
nickel, chrome and molybdenum steel mixed with small percentages of 
carbon, in their behavior towards acids and in their capacity as electric 
conductors;"" also nickel steel and nickel-iron-carbon compositions in 
their relation to compressing and shearing resistance.f 

As to the qualifications and action of nickel steel, in regard to 
extension, when heated, M. Guillaume has made many researches, 
some of which have been published.:}: 

The nickel steel used in the crank-shaft of the Noi'th German Lloyd 
steamer Kaiser FHedrich, has a tensile strength of 62 kg. per sq. mm. 
with 20.5% extension. At the break it shows fibers like those of 
wrought iron, furnishing proof that it will not break suddenly like, 
ordinary cast steel. 

Experiments have shown that a composition with 3.25% of Ni 
has 30% greater tensile strength and 75% greater elasticity than 
carbon steel. 

According to Guillaume, nickel steel with 35% of Ni will expand 
only one-tenth as much as platinum and one-twelfth as much as iron 

* Journal. Iron and Steel Institute, 1902. 
t Stahl uiid Eisen, 1902, p. 182, etc. 
i Comptes Rendus, 1897. 



350 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



Mr. Rohwer. and steel, combining the minimum factor of density with the minimum 
factor of elasticity. 

According to experiments conducted in Germany, Table 75 has 
been compiled. 

TABLE 75. 



Percentage of 
nickel. 


Expansion by heat, 
in mill, per 1» cent. 


Density per 1°, 
Celsius. 


Elastic modulus. 


0.0 
24.0 
81.4 
35.7 
44.4 
100.0 


10.3 

ir.5 
'b'.m 

8.5 
13.5 


7.84 
8.06 
8.01 
8.10 
8.13 
8.85 


23.0 
19.3 
15.3 

14.7 
16.4 
21.6 



Rivets made from nickel steel of the following composition : nickel, 
3.4%, carbon, 0.25%, manganese, 0.58%, and sulphur, 0.008%, rolled 
when red hot, show a limit of rupture of 3 400 at a strength of flexure 
of 5 900, and an extension limit of 22 per cent.* 

Guillaume found that nickel steel combines resistance against 
corrosion with a high resistance of stresses and great malleability in 
case of working. 

Nickel steel with 36% of Ni has the lowest coefficient of expansion 
known. 

With the opening of the very extensive fields of sulphur compounds 
and silicates (chief of which is nickel pyrites of iron and copper) at 
Sudbury, Canada, but recently discovered, much more nickel may be 
produced in the future than in the past. The following quantities 
have been produced : 200 tons in 1840, 4 427 in 1896, 6 898 in 1898, 
7 526 in 1900, 8 600 in 1902, and about 10 000 in 1905. 

With the increase in the use of nickel there is no doubt that its 
production will increase correspondingly; otherwise, the beneficial 
effect of its combination with iron as nickel steel in its relation to 
bridges, shown so clearly and comprehensively in this excellent paper, 
will not be realized to its fullest extent. 
Mr. Wagner. Samuel Tobias Wagner, M. Am. Soc. C. E. (by letter). — It is very 
unusual that the Society should have presented to it a single paper 
which covers so thoroughly the properties of a structural material 
about which comparatively so little has been written and for which so 
much has been claimed from time to time. Dr. Waddell is to be 
congratulated for presenting so complete a series of tests. 

The general use of a material for structural purposes, however, 

must of necessity progress but slowly, and must inspire confidence as 

it shows, in a practical manner, adaptability for its intended use. Its 

practical manufacture in commercial quantities, its adaptability to 

* Journal, Am. Soc. Naval Engrs., 1898, p. 1038; Stahl itnd Eisen, 1899, p. 1020. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 351 

withstand modern methods in shop treatment, erection, and service in Mr. Wagner, 
the structure, if successfully shown, will do much to give confidence 
and extend its use in the future. 

It took a number of years for steel to supplant iron, and the history 
of the early uses of steel for structures is one that it would be well 
to have in mind in the introduction of nickel steel. There is no use 
traveling over the same ground twice. 

The general impression which is made by a perusal of the paper is 
that for some time the use of nickel steel is likely to be confined to 
long spans, where it is possible to take advantage of its peculiar 
characteristics and to keep it under careful supervision. For structures 
of ordinary length, its use would seem to be more questionable, as the 
generally recognized grade of structural steel has many properties 
that commend it to the user. The material which we are now able to 
obtain, with an average tensile strength of 60 000 lb., will be all that 
can be desired for ordinary purposes, until we have more data relating 
to this or any new material than we have at present. 

The tests of a practical nature given in the paper seem to show 
that nickel steel of the grade tested, when compared with carbon steel 
of corresponding strength, is better than might have been expected. 
Generally, however, such tests do not show as good results as with 
the carbon steel, although in some cases the differences are very small. 
They all indicate, however, in a measure, the same relative differences 
that existed when steel was first used, and when the tendency was to 
use a high-carbon steel in order to obtain a high elastic limit and 
tensile strength. The practical manipulations of such high-carbon 
steel were not satisfactory. 

The idea, of using in ordinary structures nickel steel for a portion 
of the structure and carbon steel for the remainder, does not appeal to 
the writer, for a number of practical reasons. For long spans and for 
structures of considerable magnitude, its use in this manner is not so 
objectionable. 

The tests which have been made for corrosion are such as to leave 
the result in doubt, but, at the best, tests of this character are very 
unsatisfactory, and the only practical method of determining any 
advantage which nickel steel may have in this particular over carbon 
steel will be by experience in its use. 

At first sight it seems surprising that the coefficient of elasticity 
was not higher in the nickel steel than in the carbon steel, but a 
comparison of tests made on different grades of carbon steel seems to 
indicate that this might have been expected. Being as it is, the co- 
efficient of elasticity seems to be practically constant for all grades of 
steel, and nickel steel does not seem to offer any advantages for use in 
columns and struts. 

For material which is subject to heat treatment in shop manipula- 



353 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Wagner, tions, such as eye-bars, the utmost care must he exercised when steels 
are used which contain from 0.35 to 0.45% of carbon. Results of such 
tests are clearly sliown in the records of the tests of the eye-bars for 
the Blackwell's Island Bridge, which bars contain from 3.22 to 3.76% 
of nickel. It is a difficult matter to make a good showing with full- 
sized tests on eye-bar material which has such amounts of carbon, on 
account of difficulty in the heat treatment. There may not be very 
great differences in the ultimate strength, but the variations in the 
elongation and the character of the fracture are likely to be trouble- 
some to the manufacturer. Table 60 gives the list of tests that met the 
specifications. It would be very interesting to know what difficulty 
was experienced, if any, in meeting the specifications on this large 
work. It is this feature of the case that appeals to the writer, 
especially in the use of such material. If no special difficulty was 
encountered, the resulting product would be reasonably sure to be 
satisfactory as a "wjhole; if there was a struggle to obtain the result, 
some not altogether satisfactory material is likely to get into the work, 
although it may not be shown by the tests. Uniformity in the character 
of the product goes a great way toward good service in the structure, 
and when a manufacturer is called upon to produce some special 
grade of material which is out of the ordinary run of his practice, the 
greatest care must be exercised in order to accomplish the desired 
result. 

Some years ago, when structural steel was in its infancy, the writer 
was inspecting some 60 000-lb. steel at a plant where this grade of 
material had never been made before. The orders were small, and, 
between the heats which were made for him, heats of high-carbon steel 
were manufactured in the same furnaces. For a long time, whenever 
an attempt was made to change from the high-carbon to the low- 
carbon steel, the first heats were never quite right. After considerable 
experience, this difficulty was overcome. In changing from a low- 
carbon steel to a high-carbon nickel steel, the same difficulty would be 
looked for, to a certain extent, in any existing steel mill, and for a 
while would probably interfere with the thoroughly uniform character 
of the nickel-steel product. Time and experience, of course, would 
reduce this difficulty to a minimum. 

Wliile the tests that have been presented are unusual in their 
extent, many more data will be needed before the use of such steel will 
be specified for spans of ordinary length. The thanks of the Profession, 
however, are due to Dr. Waddell and his associates for making this 
mass of valuable information public, and it is to be hoped that the 
discussion will bring forth information from those who have already 
used nickel steel for structural purposes. 
Mr. Carpenter. A. W. Carpentkr, M. Am. Soc. C. E. (by letter). — The writer 
wishes to express his admiration for the scope and thoroughness of 



DISCUSSION ON NICKEL STEEL FOR BKIDGES 



353 



Dr. Waddell's investigation, and to acknowledge his indebtedness Mr. Carpenter, 
for access to such a treasure of test data as has been disclosed. 
To gather the information and data required for this paper must have 
been an enormous task and expense. The result cannot fail to add 
to Dr. Waddell's already great fame as an engineer and investigator. 

There would seem to be little doubt left as to the adaptability of 
nickel steel for bridges, and its economy over carbon steel for long- 
span bridges — say, spans of more than 500 ft. For railroad bridges of 
the ordinary range of span, the writer has not been convinced that it 
is desirable to change from carbon steel to nickel steel. In the first , 

place, he would suggest that the pound prices, given as averages for 
different classes of bridges on page 153, are a little high for the first 
three classes, at least when the location is in the States of New York, 
Pennsylvania, or Ohio, and would suggest the following revision, 
bearing in mind that the erection cost should not include in any way 
the cost of supporting traffic nor of dismantling old structures to be 
replaced, which items form a large portion of the cost of erecting 
railroad bridges: 

Deck, plate-girder spans 3.5 cents. 

Through, plate-girder spans 3.75 " 

Riveted and pin-connected Pratt-truss spans. . 4.00 " 

Accepting Dr. Waddell's assumptions otherwise, his Table 6 would 
be modified as shown in Table 76. 

TABLE 76.— Percentages of Excess Cost of Carbon-Steel 
Bridges over Mixed-Steel Bridges. 



Type of Structure. 



Single-track, deck, plate-girder spans 

Single-track, through, plate-girder 
spans 

Single-track, through, riveted, Pratt- 
truss spans 

Single-track, through, pin-connected. 
Pratt-truss spans 

Double-track, through, riveted Pratt- 
truss spans 

Double- track, through, p i n - c o n - 
nected, Pratt-truss spans 



Span Limits, 
in Feet. 



20 to 120 
30 to 110 
120 to 200 
190 to 310 
110 to 180 
180 to 310 



Least 
Excess. 



— B.5 
+ 1.5 
-f-1.4 

0.0 
-h 0.9 

— 0.6 



Greatest 
Excess. 



4-7 
+ 8 
+ 4A 
-f 5.8 
-f 5.0 
+ 6.3 



Approximate 
Average. 



— 1.5 

+ 5. 
+ 3. 
+ 3. 
-f 3. 
+ 3.5 



Table 76 shows that deck, plate-girder spans of "mixed" steel would 
cost, on the average, more than similar spans of carbon steel; and that, 
for other types, the "mixed" steel construction would be generally a 
trifle cheaper. The writer does not understand that the author claims 
or demonstrates that a nickel-steel structure would be preferable to 
one of carbon steel of equal strength, at equal cost. Therefore, he 
sees no argument for the use of nickel-steel, deck, plate-girder spans, 
and feels that there are good arguments for the use of carbon steel 



354 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Carpenter, in bridges of the other types given in Table 76, even at the greatest 
excess cost shown. Some of the points which appear to argue in favor 
of the carbon steel are tlie following: 

Experience. — The reliability and serviceability of carbon steel has 
been demonstrated by years of the most extensive experience. There 
is a possibility that, in spite of Dr. Waddell's most thorough investiga- 
tion, some seriously detrimental quality in the nickel steel might 
develop between the mill and the finished structure ten years or more 
in service, which it would not be worth while to risk for the small 
saving in first cost. 

Availability.— There seems to be no doubt that it would take longer 
at this period to fill an order for nickel steel than one for carbon 
steel. Even with nickel steel in general use, the addition of another 
grade of commercial steel would no doubt result in increasing the 
average time for filling material orders. 

Danger of Mixing Steels. — The only way to effect economy in the 
use of nickel steel for ordinary spans appears to be in the combined 
use of nickel steel for parts proportioned by unit stresses which require 
more material than the specified minimxim sections in carbon steel 
would supply and of carbon steel for parts in which the minimum 
sections would furnish sufficient strength, including ordinary laterals, 
stiffeners, fillers, base-plates, etc. On page 248 the author states: 
"Extreme precautions were taken at every step to keep the two steels 
separate." As the steels cannot be distinguished by surface appear- 
ances, it would seem that the use of "mixed" steel would be attended 
by much trouble in distinguishing the two kinds, and in danger of 
getting them interchanged. A mistake in a main member might be 
disastrous. 

Elastic Distortion. — On page 150 the author refers to the greater 
deflections of nickel-steel spans over those of carbon steel, but thinks 
this not serious. The matter of stiffness, however, has been given 
sufficient serious attention to warrant fixing the limits of depth to 
length ratios of trusses and girders in prominent specifications. Such 
specifications require that, in case the limits are exceeded, the sections 
shall be increased so that the deflections shall not exceed that of spans 
of normal sections and limiting depth; and these specifications apply 
to structures designed for carbon-steel unit stresses. With nickel steel 
having the same coefficient of elasticity as carbon steel, but stressed 
60% higher, the deflection under live load would be about in proportion 
to the unit stresses. If there is any good reason in the usual specifica- 
tion for limiting depth ratios for carbon-steel structures, then there 
is as good reason to make the depth limit for nickel-steel structures 
such as would give equal deflection, or increase the sections to give 
the same result. If the limiting depth for a carbon-steel girder is 
one-twelftli of the span, then that for a nickel-steel girder of equa' 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 355 

strength would be a much larger ratio, say between one-ninth and Mr. Carpenter. 

one-tenth of the span. This factor would especially affect long girder 

spans, and probably was not considered by Dr. Waddell in preparing 

his tables of weights and costs. There are a number of other ways ' 

in which the comparatively great elastic movement of nickel steel, 

strained with the author's working stresses, might cause trouble, and 

these he does not appear to mention. One of these is the extension 

of floor systems in which the stringers are riveted into the floor-beam 

webs. Assuming a 25-ft. panel and nickel-steel chords, with a live-load 

stress of 18 000 lb. per sq. in., and carbon-steel chords with 10 000 lb. 

per sq. in. (about the author's working ratios for probable actual 

stresses). 

The carbon-steel chords would extend 0.10 in. 

And the nickel-steel chords 0.18 " 

Would the riveted stringer connections stand the strain with the 
nickel steel? What would be the lateral bending stresses on floor 
beams? Again, in riveted-truss spans of equal depth, would not the 
nearly doubled deflection of the nickel-steel span greatly increase 
the secondary stresses in the riveted joints, which, if properly com- 
pensated for, would wipe out all the economy of nickel steel? 

Again, what would be the effect of the excessive stretching and com- 
pressing on paint and preservative coatings? The elastic distortion 
due to the dead load could no doubt be provided for much better than 
that due to the live load, as the former would be a fixed quantity for 
any span after being swung, while the latter would vary from zero 
to a maximum and back, with the passage of every live load. With 
very long spans, the ratio of dead load to live load becomes large, and 
reaches a point at which the distortion of main members due to live 
load would be no greater than in carbon-steel bridges of short span. 
At and beyond such point the factor of elastic distortion would be 
mainly a dead-load matter, and could no doubt be easily provided for. 

A point in connection with the properties of nickel steel which the 
author has not brought out quite to the writer's entire satisfaction, is 
the effect upon it of cold-straightening. The writer has seen some 
pretty bad kinks, especially in long, thin, girder cover-plates, which 
have been taken out in cold-rolling, and wonders if nickel steel would 
be as little affected by such treatment as carbon steel. The impact 
tests, described on page 107, which did not show to the advantage of 
the nickel steel, were of this nature. 

The proposition to use larger rivets for nickel steel would be rather 
objectionable for field riveting. One of the great advantages of the 
nickel steel would seem to lie in the shorter rivet grips required on 
account of the thinner metal that would be used than with carbon 
steel. This alone should make rivets in nickel steel more efficient. 



356 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Carpenter. It might be pertinent here to mention that there are other alloy 
steels which appear to have even more remarkable properties of strength 
and ductility than nickel steel, and probably to cost no more. Some 
of these have been developed in the automobile industry. They would 
seem worth investigating prior to the construction of another bridge 
of the East River or Quebec Bridge proportions. The author has 
shown how such an investigation should be made, and a similar paper 
on vanadium steel for long-span bridges would be valuable. 

The test data furnished by the author in the appendices seem to 
the writer wonderfully complete and valuable. Some of the results 
are surprising, especially the low values for elastic limit and yield 
point shovpn in many cases. It would seem that a limit should be 
placed on the speed of testing machines, in ordinary practice, for 
the detennination of the yield point. The writer would like to know 
whether the author considers it practicable to insist, in ordinary prac- 
tice, that the determination be made with the beam balancing, and if 
this will insure the proper low speed of test. 

Regarding the tests of struts, the writer is rather disappointed that 
so little information is given as to the accuracy of the readings from 
the Phoenix machine. The calibration of this machine in connection 
with the cast-iron column tests conducted by the New York City 
Department of Buildings in 1897 showed that a correction of some 
15% to the readings of the gauge on the machine was necessary to 
reduce the same to the actual pressures on the columns.* The author 
states that the figures given (for loads presumably) "are based on 
data furnished by the Phoenix Iron Company as to the relation existing 
between the mercury column and the pressure on the piston of the 
machine." This does not appear to take recognition of any frictional 
losses in the piston. It is hoped that, in view of the present wide- 
spread interest in the strength of columns, the author will take up this 
point more fully and remove the doubt that now exists regarding the 
reliability of the results he records. The suspicion that the loads 
recorded for the corresponding compressions are too high is somewhat 
confirmed by calculating the loads from the compressions, using the 
factor determined by the New York City Building Department tests. 
These tests showed that a compression of 0.001 in. in 200 in. in a 
Phoenix colunni (soft steel) 14i in. inside diameter, 21 ft, long, 75.3 
sq. in. section, was produced by a load of 149 lb. per sq. in., as 
measured by the Watertown Arsenal testing machine, the accuracy 
of which is unquestioned. Applying this factor to the compressions 
of the 10-ft. struts on a lengtli of 90 in., the results are as shown in 
Table 77. 

* Engineering News, Jan. 18th, 1898. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



357 



=5 s 

(tt p> 
O HcnO 

o =■ r o 

m'^ 3 



-,» 

ft> o 

SB 
g,cc 



O-B 



£3 
5' 3 



So 



P" in- 



JT "O 






ooo o coooo oo 
ooo ts ooooce cncc 



•-^oc 1*^ o 

88 8 



88888 88 



CdlOCD *<) I CO O Oi CO 4^ O^-Q 

OQO O O Q O O- O QO 

OOO Oiooooo oo 



o ooo 
00 ooco 



«S OS CO—OD 

cc) CO c:»o O 

8 8 882 



lb. ik CO CO 

-I — a3 »j 



_op 
oo 

•t> 00 

ooo 



Compression, 
in inches. 



Load by 
Table 51, in 
pounds per 
square inch. 



Calculated 

load, in 
pounds per 
square inch. 



Compression, 
in inches. 



Load by- 
Table 51, in 
pounds per 
square inch. 



Calculated 

load, in 
pounds per 
square inch. 



Compression, 
in inches. 



Load by 
Table 51. in 
pounds per 
square inch. 



Calculated 

load, in 
pounds per 
square inch. 



Compression, 
in inches. 



tOK) Kl>-' 



oc o o 
coco 



Load by 
Table 51, in 
pounds per 
square inch. 



too rf*C5 
C7TO O O 

So oS 



SSa 



oo CO 
CO c'c 



COCOtOtO {3t0 iOl-' 

COl-i^OS Vf <t^ K)^ 

Sgg? 5c gS 

ooccoo CO 



Calculated 

load, in 
pounds per 
square inch. 



Compression, 
in inches. 



Load by 
Table 51, in 
pounds per 
square inch. 



05ift.>-'0' « OS •t' N) 

wocootn cccn •.=■•-■? 

8888188 88 



Calculated 

load, in 
pounds per 
square inch. 







H 






> 




„ 


bd 


•y. 


M 


o 


■^ 


W 


1^ 


tH 


^T 


H 




-~I 




O 


1 


^ 


Q 


o 

o 




■<' 


H^ 


H 




T) 




'^ 


?0 


H 


w 


ffJ 


O 


W 


^ 


Kj 


O 


V, 


H 


>. 






Ui 




H 


o 



2 i^; 



> 
td 



M 


O 


H 




> 


Or 


'■ ■> 


f 


M 


< 


o 

c 


*1 


> 


> 

H 


O 


O 












i^ 


t-l 


M 


►n 


u 


o 


CO 

r/i 


O 
w 


o 


O 


O 


H 
H 


o 




Wl 


O 


H 


5 


o 




> 


o 

5 


Q 

> 

s 


o 




w 


H 


W 


> 


w 


.o 
n 


o 


o 


>-' 




CO 


w 


o 


o 




"^ 




w 




t— 1 


!> 


H 


o 


M 


W 


a 


o 

> 


hd 


jn 


o 


»H 


> 




H 


^2! 


11 


!2| 


O 



>< 9 



O CO 



Mr. Carpenter. 



358 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Moisseiff. Leon S. Moisseiff, M. Am. Soc. C. E. (by letter). — To one 
familiar with the scantiness of information on structural nickel steel 
available in any language, the great value of Mr. Waddell's paper is 
apparent. Several papers, based on a few laboratory experiments, 
comprised all the published information until a few years ago. This 
was supplemented by some semi-confidential replies to inquiries from 
mills and producers of nickel, which were based more on opinions 
and hopes than on demonstrated facts. 

When, in 1902, or about a year before Mr. Waddell began his 
experiments, Gustav Lindenthal, M. Am. Soc. C. E., then Commis- 
sioner of Bridges of New York City, wished to utilize the greater 
strength of nickel steel in the design of chains and stiffening trusses 
for the proposed Manhattan Bridge, and partly for the eye-bars of 
the Queensboro (then Blackwell's Island) Bridge, this was the state 
of affairs: At the request of the Department of Bridges, some melts 
of nickel steel were then made, and a few eye-bars were tested. The 
information thus obtained served as a basis for the Queensboro Bridge 
specifications, which covered eye-bars and pins only. Judging from 
the results of the eye-bar tests, as contained in the records of the 
Department of Bridges, these specifications have, with some minor 
changes, produced very good material. 

The lack of information on the behavior of nickel steel in the 
various processes of fabrication, its effect on the cost of such fabrica- 
tion and on the efficiency of the tests in use in bridge shops, was felt 
by engineers who wished to avail themselves of the new material, and 
by the producer who desired to bring it into the market. It was to 
supply the necessary information to the engineering world that, the 
writer believes, Mr. Waddell began his series of experiments. 

The researches, as planned by the author, were well laid out to 
cover most of the information needed for the designing of bridges 
and the writing of modern bridge specifications. It is to be regretted 
that the intended series of "small special melts of nickel steel con- 
taining varying proportions of nickel, carbon, and possibly manganese" 
was not carried out. Such a series of melts would have settled, for 
some time at least, the question of the best composition of nickel 
steel, and would have much strengthened the author's conclusions. It 
will be comparatively easy for the designing engineer to get now and 
then some special tests on points of fabrication, or on the efiiciency of 
some details of nickel-steel members, but he will encounter difiiculties 
when asking for special melts, and will sometimes have to follow 
advice which is not always disinterested. 

It is not, of course, to be expected that any single paper, in a 
virgin field such as that of nickel-steel tests, can cover all points and 
satisfy all inquiries. In due time special tests will, presumably, be 
made of one or the other qualities and applications of the new 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 359 

material as it comes into use, but only the real test of experience will Mr. Moisselff. 
decide matters fully. Meantime, this paper furnishes enough informa- 
tion to enable the engineer in need of a high-resistance steel for 
structural purposes, to utilize the advantages of nickel steel within 
safe limits. 

Whether or not one will agree with the exact composition of the 
nickel steel proposed by the author, or with the working stresses 
recommended by him, it must be admitted that he has created a work- 
ing basis, both for the design of nickel-steel bridges and for further 
experiments. His comparative cost diagrams demonstrate at sight 
the economy of nickel-steel bridges within the extra charges for this 
steel computed by him. It is quite possible that a heavy demand for 
this material may raise the price of nickel above the estimated extra 
charge of 2 cents per lb. of nickel steel, but the increase of cost would 
check itself automatically by the decrease in demand. A glance at 
the cost diagrams for the longer spans, however, shows that, as may 
be expected, even a considerably higher extra charge would still be 
found to be economical. 

It has been the writei^'s fortune to make use of nickel steel in three 
cases, each of which differed essentially in the functions to be per- 
formed by the members, and also in the reasons for adopting nickel 
steel as the proper material for them. To the writer's knowledge, 
these are the only instances, up to the present time, of the use of 
nickel steel in main members of bridges. It also happens that these 
three instances are fairly illustrative of the conditions which may 
lead to the use of nickel steel, even aside from economical con- 
siderations. 

The Queensboro Bridge was the first in which nickel-steel members 
were used. The great weight and capacity of this bridge required in 
the chords sectional areas of unprecedented dimensions. At that time 
it was not thought desirable to demand from manufacturers eye-bars 
wider than 16 in., and the excessive width of the tension chords, 
together with the long pins, was the primary cause which led to the 
adoption of nickel-steel eye-bars and pins. The resulting economy, 
while making the attempt desirable, was not deciding in this case. 
The vpriter's connection with the design of this bridge extends only as 
far as the original contract design. 

The Manhattan Bridge affords a more important example of the 
advantages of a high-resistance steel than the preceding. This bridge 
will have the greatest capacity of any long-span bridge yet built. It is 
designed for four rapid-transit tracks, four surface-car tracks, a 
35-ft. roadway, and two footwalks, each about 11 ft. in width. This 
means practically fourteen lines of trafBc. For this traffic a "working" 
load of 8 000 lb. and a "congested" load of 16 000 lb. per lin. ft. of 
bridge have been allowed in the computations. Under "congested" load 



360 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Moisseiff. is meant a moving load of maximum density, or as dense as the 
dimensions of cars and vehicles will permit. The "working load" has 
been assumed as one-half the "congested" load; or, in other words, it 
has been asstimed that trains, cars, and vehicles will have to operate 
within their own distance apart. It is well known that for a practically 
uniformly distributed load over the entire bridge, the stresses in the 
stiffening trusses will be, as a rule, comparatively low, and that they 
will attain their greatest values for certain partial positions of load- 
ing. To get the greatest stresses in the stiffening trusses, the fourteen 
lines of traffic of this bridge would all have to stop at the theoretical 
load limits, and the temperature would have to be simultaneously at 
one of the extreme deviations from the "normal" assumed for it. The 
probability of ever realizing this condition is very remote indeed. But, 
in a structure of such magnitude and importance, it is good engineer- 
ing to provide for any possible, if even not probable, condition, so 
that the structure may be then not only safe but uncrippled. Propor- 
tionately high unit stresses may then be allowed for such greatest 
"congested" load stresses. 

To take care structurally of the large stresses deduced from the 
above extreme conditions, the stiffening. trusses would have to be deep 
and unsightly, or contain very heavy chords. In either case, the 
moment of inertia of the trusses would be increased, which would in 
turn increase the stresses, both due to temperature and to moving load, 
without any appreciable benefit to the cables and towers. It thus 
became apparent that stiffening trusses of a limited moment of inertia 
are desirable, and that only a high-resistance steel, such as nickel 
steel, will satisfy the requirements. Here, again, the economy to be 
realized in the cables and towers was an additional inducement to 
adopt nickel steel, there being more than 8 000 tons of it, but the 
primary cause, as previously explained, was the need of a reliable 
high-resistance steel which would usually be stressed low but which 
might possibly need at some time its full strength. 

The third instance of the use of nickel steel was in connection 
with certain reinforcing work on the Williamsburgh Bridge. The 
writer had found that, due to the increase in weights of cars and other 
reasons, the rocker posts at the ends of the stiffening trusses and their 
supports required strengthening. The necessary supports were pro- 
vided by the erection of what is practically a fifth leg in the main 
towers. The rocker posts, however, had to be of the same dimensions 
as the existing ones in order to fit in the rather cnxmped space. It 
was a simple thing to replace thoui by nickel-steel rockers and pins 
of almost identical dimensions, and to increase thereby their strength 
by 50 per cent. In this case the cost of the material did not enter 
into the consideration at all. 

The writer believes that those three instances demonstrate fully the 



DISCUSSION ON NICKEL STEEL EOR BRIDGES 361 

desirability of a reliable structural steel of high resistance in many Mr. Moisseiff. 
important cases, aside from economical considerations. 

The writer docs not agree with Mr. Waddell as to the desirability 
of making the floor systems of bridges of nickel steel. The design of 
the floor system is frequently determined by other considerations than 
mere theoretical strength. The stiffness of beams thus often determines 
their dimensions, and minimum sections are not infrequent, even in 
what is known as "structural steel" beams. Buckle-plates, guard-rails, 
etc., will remain of the same weight. It certainly will require cheap 
nickel steel to realize much actual economy in nickel-steel floors. 

James C. Hallsted, M. Am. Soc. C. E. (by letter). — The writer Mr. Haiisted. 
is very much impressed with the large amount of labor Mr. Waddell 
has performed in gathering information and preparing it for this 
paper. He has certainly added a great deal of very valuable informa- 
tion on nickel steel, and deserves the thanks of the Engineering Pro- 
fession. He has demonstrated the advantages of nickel steel in eye- 
bars, for heavy spans, and has pointed out the possibility of its 
economic xise in heavy compression members. 

The writer is not as sanguine as the author as to the advisability of 
using nickel steel in ordinary bridge work. He has clearly shown the 
difficulties in the way of its use, but, in the writer's judgment, he has 
underrated their importance. 

A sample order could readily be made by the mills, in some such 
way as they can turn out rails and ship them to a foreign country at 
a less rate than they sell them for domestic use. The cost of this 
sample order is not necessarily a criterion of the cost that would pre- 
vail if nickel steel were to be made as a general thing in the mills. 

The wear and tear on the equipment of a rolling mill in rolling 
the harder steel would not be evident in rolling a sample or an 
occasional order, as would be the case if the mills were constantly 
rolling nickel steel. 

Further, the comparative cost of making eye-bars in nickel and in 
carbon steel is not representative of general structural work. Large 
forgings do not tax the wearing quality of the equipment like punch- 
ing, shearing, reaming, and drilling; and the shopwork on eye-bars 
is much less than on built members. A difference of 1.5 cents per lb. 
for the two grades of steel does not appear to be enough, especially 
"when 1 cent of this is for the raw nickel. Entering into the use of 
nickel steel there are many factors which would increase the cost to 
more than that of carbon steel. Some of these will be mentioned. 

There are apt to be many rejections at the mills, not only because 
nickel steel is a new material and demands more expert treatment, but 
also because the requirements must be strictly adhered to if the high 
units are to be justified. Comparatively high rejections mean greater 



362 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Halisted. cost; they also mean scrap. To use nickel scrap economically, there 
must be created a demand for nickel-steel castings, and there are in 
structures few places where nickel-steel castings are required. 

The rolling of nickel-steel shapes has scarcely been tried. The 
ramifications of the various shapes make their rolling more difficult 
in hard steel than in soft steel. "Wear or breakage on equipment, 
reduced output, imperfect shapes, all mean increased cost. The thin 
webs of standard channels would be especially troublesome in the 
finishing passes; and if the thin-webbed channels be avoided for 
heavier weights, economy is lost. 

Pimching is harder on the tools in nickel steel than in carbon 
steel. ]\fr. Waddell suggests the use of larger rivets, so that the holes, 
being of larger diameter, will be less destructive to the punching tools. 
This could scarcely be done in the flanges of channels, as the standard 
holes now used are generally the maximum for safety. In the general 
run of built members made of channels, the holes are nearly all in the 
flanges. Reamed work is a practical necessity in nickel steel. Sub- 
punching in channel flanges is scarcely a possibility, because of the 
small size of the holes. If shops are driven to drilling, with the hard 
usage that drills will get in nickel steel, shop cost will go up by leaps 
and bounds. 

Larger rivet holes in tension members mean greater reduction of 
net section, another dash in the scale pan which will weigh against the 
economy of nickel steel. 

In the matter of inspection and tests, costs will also run up. It 
would be harder to keep track of nickel-steel rollings among a lot of 
carbon-steel rollings of the same shapes, and greater care would have 
to be exercised in separating the nickel steel, because of the menace 
to a structure which would result in using carbon steel where nickel 
steel was intended. 

Even in a mill where only nickel steel is manufactured, the cost of 
tests made according to Mr. Waddell's proposed specifications would 
be great. For example, the elastic limit is the load which produces 
a permanent set of 0.01 in. in 8 in. To make this determination the 
machine would have to run very slowly; it would have to be stopped 
and reversed many times near the elastic limit in order to see whether 
the specimen had a permanent set of 0.01 in.; very careful measure- 
ment would be required to detx^rmine just when this 0.01 in. had been' 
reached. The time consumed in making the determination of the 
elastic limit on one nickel-steel test, by this method, would suffice to 
make several complete tests on carbon steel. To stop or reverse the 
machine in testing eye-bars would also add to the cost and trouble. 

There is no doubt that a manufacturer will add to his price on a 
job on which there is inspection, especially if he thinks it will be 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 363 

rigid, and it would be dangerous to use nickel steel without rigid Mr. Haiisted 
inspection. Much as inspection is desirable on all work, the fact can- 
not be ignored that a large part of the output of the shops is not 
inspected. Work of this class would scarcely be attempted in nickel 
steel, so that there would always be stock angles or beams of carbon 
steel lying around, which it might be "necessary" to use in a nickel- 
steel structure, if the ordered material were delayed or defective. 

Mr. Waddell has shown that, with nickel steel, all the shop pro- 
cesses are more expensive in labor, require better tools, and are harder 
on tools, than if carbon steel be used. It would appear that better 
shop equipment and special shops will be required, if much nickel- 
steel work is to be attempted. All these things mean higher cost, and 
seem to mean it to the extent of more than 0.5 cent per lb. 

As an example of the difficulties and expense that would result in a 
general use of nickel steel in riveted work, Mr. Waddell's statement on 
page 239 will be quoted: 

"The speed, therefore, with which the carbon steel was drilled 
was 1 in. in 35 sec., and the nickel steel, 1 in. in 52 sec. A blue-chip 
tool, in ordinary work, would last half a day without sharpening, 
whereas, if used for 5 or 6 min. on nickel steel, it is necessary to 
sharpen it." 

There can hardly be said to be a demand for a material having 
the qualities of nickel steel, except in rare cases. The one feature in 
which nickel steel excels carbon steel is in its higher elastic limit and 
ultimate streng-th, with the possible exception that it appears to stand 
weather somewhat better. Nickel steel is not as ductile, not as tough, 
not as resilient as carbon steel; it requires more careful and more 
skilled treatment, and; cannot stand the punishment that carbon 
steel will. 

The relative dead weight of a nickel-st-eel and a carbon-steel struc- 
ture is so small that there would be little difference in the stresses of 
an ordinary span. There are fixed weights, such as the floor, the 
lattice bars, tie-plates, etc., which would be the same in either, so that 
the relative sectional areas would not represent the comparative 
weights. There can be little saving, therefore, in the weight of a 
structure of ordinary dimensions, if, for the time being, we lay aside 
the diminished sections due to higher unit stresses. It can scarcely 
be said, then, that a demand for nickel steel exists, based on the saving 
in weight, from the standpoint of stress in members for ordinary 
structures. 

In very long spans the dead weight of truss members is a large 
factor in the stresses on those members, and anything that will reduce 
that dead weight is economically advantageous. In Europe, some 
years ago, a suspension bridge was built with chains made of eye-bars 
cut out of high-steel plates. Nickel-steel eye-bars would have met the 



364 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Hallst€d. demand created by the design of this bridge most admirably, and 
would have meant a large saving in expense. Nickel-steel eye-bars 
would be economical, no doubt, in spans of moderate lengths, if the 
tonnage justified a special order; and it may come into general use 
for eye-bars for all spans. 

In compression members, nickel steel would possibly be economical 
for heavy short members. In these, direct compression plays a large 
part, and bending stresses are almost eliminated. Slender members 
in hard and soft steel approach the same ultimate strength as the 
ratio of length to radius of gyration increases. It can be shown that 
the load of Euler's formula, instead of being the compression which 
a column may sustain at any deflection within the elastic limit, as 
commonly stated, is actually the absolute maximum load that the 
column will carry. This load will produce failure in a column, no 
matter how high the elastic limit. This may appear startling. For 
proof the writer refers to a paper on the subject by Mr. Edward 
Godfrey.* This being the case, slender columns in nickel steel will 
approach equality in compressive strength with carbon-steel columns, 
so that for slender members the relative advantage of nickel steel 
becomes still less. It is not an accident that Mr. Waddell's tests on 

columns in nickel steel at - == 27 averaged 75% greater in ultimate 
strength than similar columns in carbon steel, while those on nickel- 
steel columns at — =: 81 averaged only 47% stronger. The writer does 

^ * I . . 

not understand, however, why Mr. Waddell uses 30 000 — 120 in his 

r 

compression chords, in view of the facts shown by his tests. This, as 

compared with the carbon-steel formula (IG 000 — 70 ^ ), gives an 

r 

allowed load 90% greater at 27 radii and 96% at 81 radii, against 75% 

and 47%, respectively, as found by test. 

The writer thinks that Mr. Waddell's base of 30 000 in the com- 
pression formula is too high. It is a factor of about 2.6 at 27 radii, 
and of only 2.2 at 81 radii, according to his tests. The factor of 
safety for tension members is more than 3, and tension members are 
not affected by imperfect shopwork or alignment in anything like the 
amounts suffered by compression members. 

If a truss in nickel steel is made of more slender compression 
members than one in carbon steel, these slender members will approach 
nearer the point where carbon steel and nickel steel have equal strength, 
and the economy of nickel steel tends to vanish. 

The impact test of materials has been developed with the idea of 
demonstrating two qualities of a material. These are, first, to ascer- 

* To be publisbed^soon in the Railroad Age Gazette. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 365 

tain whctlier the material is in an abnormal or dangerous state, that Mr. Halisted. 
is, whether or not it is brittle; and second, to determine its resistance 
to shock under working conditions. 

Impact tests have been conducted in one of two ways, either by 
the single-blow method or by the repeated-blow metliod. Both these 
methods of testing show unquestionably whether or not the material 
is brittle. 

The repeated-blow tests are designed to determine the resistance 
of tlie material to shock under conditions which will approximate the 
actual conditions of service. In the writer's opinion, they do not 
furnish this information. All repeated-blow tests, of which he has 
knowledge, have been essentially to test the toughness of the material 
rather than its ability to resist repeated blows. In every test, each of 
the repeated blows has stressed the material much beyond the elastic 
limit or yield point, and the second and third blows, whether upon the 
same side of the test bar or upon the reversed side, have simply 
carried the deformation further, until ultimately the material failed 
under the test more as a result of the lack of ductility than any- 
thing else. 

Much better would be a vibratory test such as has been made on 
stay-bolt iron for many years. In such tests the plastic deformation 
of the material is comparatively slight, and, for this reason alone, the 
test would more nearly demonstrate the comparative value of two 
materials in service. 

In November, 1908, a great many data regarding impact tests were 
presented before the Institute of Mechanical Engineers. These tests 
were made on the various impact testing machines which have been 
on the market in recent years, machines in which every endeavor has 
been made to obtain as great a refinement as possible. Those used 
were the Seaton and Jude, the Fremont, the Izod, the Kirkaldy hori- 
zontal, and the Kirkaldy vertical. A careful series of tests was made 
with the idea of comparing the action of these various machines. In 
most cases it was foimd impossible to get any consistent results, and, 
worse than that, it was found that no one of the machines gave con- 
sistent results, even on identical specimens of steel. On one machine 
the results varied from 4 to 60% and more. 

The irregularities disclosed by the different methods of testing 
have l>een found to be due not to the lack of uniformity in the 
material, as was suggested by some of the investigators, but largely 
to defects in the method of testing, and it seems reasonable to con- 
clude that the method of making impact tests and the results obtained 
should not be relied on by engineers to differentiate between the 
physical properties of different materials. This conclusion was reached 
after careful tests on the various types of machines mentioned. From 
these considerations, it seems to the writer that the impact test is 



Mr. Hallstecl. 



Mr. Arnodin 



366 DISCUSSION ON NICKEL STEEL FOE BRIDGES 

only of value where a rapid method of discovering the brittleness of 
material is desired. 

In Mr, Waddell's paper, it is not surprising that inconsistent 
results were obtained on a crude apparatus, when it is considered that 
even on the most refined machine the same inconsistencies occur. The 
writer would urgently recommend that something be done in the way 
of a vibratory test. 

It is his opinion that reaming a rivet hole slightly, or planing a 
sheared edge a very small amount, say :} in., does not remove the effect 
of punching or shearing, though, without question, much good is done 
thereby. Some tests made by the writer several years ago on steel 
1 in. thick and having a tensile strength of 70 000 to 80 000 lb., went 
to show that shearing lessened the ductility for about 3 in. from 
the edge. 

As to the absence of high phosphorus, the bending test is not 
conclusive. The writer has made beautiful bends with 0.16% phos- 
phorus. As with the drift test, a great deal depends on the operator. 

In conclusion, the writer has stated his ideas on the use of nickel 
steel, not with the intention of detracting from Mr. Waddell's excellent 
paper, but to throw another light on the subject. He has emphasized 
chiefly the features affecting the use of nickel steel as they appear to 
one whose daily concern is in gauging and passing on the excellence of 
the work in the shops and mills. Only time and the developments of 
the qualities and uses of steel in tools will determine the place of 
nickel steel in structural engineering. To the writer it appears that 
tools, as made at present, are not capable of manipulating economically 
a metal having the properties of nickel steel. 

F. Arnodin, Esq. (by letter). — Whatever individual opinion may 
be entertained on the subject of nickel steel and its use in construc- 
tion, it must be recognized that this paper, with the numerous experi- 
ments it records, constitutes an important treatise, upon which the 
author is to be congratulated. 

Mr. Waddell discusses mainly the determination of the proportions 
of nickel and carbon which will produce a steel of the greatest 
strength, combined with the highest elastic limit, and a malleability 
enabling the metal to stand easily the stress of shop manipulation. 
In order to achieve this three-fold purpose, he has furnished a large 
number of analyses and comparative tests, from which the chemical 
and physical properties of the material may be determined with the 
greatest possible accuracy. 

This paper, then, is a laboratory work, and the writer, not being 
an expert in this branch of the subject, will leave the discussion of 
such features to metallurgical and chemical engineers. 

However, inasmuch as the paper considers the construction of great 
metallic structures, especially bridges, the problem does not depend 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 367 

entirely on the qualities of the steel to be used, because the kind of Mr. Arno«iin. 
members, their connection, and the manner of stressing the metal, 
whatever the latter may be, are of equal importance. Only by har- 
monizing these with one another, and with the qualities of the metal, 
can the safety, the economy, and the progress desired be realized. 
It is impossible to attain these ends, if each of these considerations is 
treated apart from the others. 

Accordingly, this contribution to the discussion is that of an 
engineer and constructor who considers the use of nickel steel in bridges 
from the standpoint of the safety, economy, and durability of the 
structure. 

Disadvantages Due to High Resisting Capacity. — At the outset, it 
should be stated that it may be dangerous to look for high resistance 
in the metal used for bridges, for, in order to profit by this high 
resistance, and build structures of proportionally lighter weight, there 
will be many members having small cross-sections compared with their 
lengths; and therefore they cannot well resist the stress due to 
column flexure which is so dangerous in metallic structures. Such light 
sections, furthermore, are at a disadvantage with respect to oxidation, 
which attacks the lighter member with the same intensity as the larger 
one, and, consequently, will ruin the former far more rapidly than 
the latter. 

Again, a light structure of a given span and strength of material 
will be influenced more by the moving load — the load of heavy trucks, 
locomotives, trains of cars, etc. — than a heavier structure with mem- 
bers of greater cross-sections. This is expressed quite neatly by the 
following figure: "The fly, when it alights upon a spider's web, causes 
more dangerous deformations than when it walks upon a table." 

In short-span bridges the deformations are small and have very 
little effect on the structure, but they acquire a dangerous importance 
in long-span bridges, especially when the form of the structure is 
such that these deformations induce secondary stresses, which are 
often difficult to analyze, and which the regular static computations 
do not take into account. 

Again, the engineer who has at heart the future of his structure, 
must, from the very inception of the design, consider the proper 
relation between the fixed loads of the structure and the traffic it 
has to carry. It is self-evident that, other things being equal, this 
relation will be more favorable with the less resistant carbon steel, 
than with the stronger nickel steel. 

There is no doubt that, with nickel steel, one may build as heavily 
as with carbon steel, and, accordingly, gain greater safety. From 
this point of view, there is no doubt that, if the cost of each material 
was the same in all cases, the preference would have to be given to 
nickel steel, the superior qualities of which have been so well demon- 



368 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Amodin. strated by the author's experiments. But the prices are different, and, 
in order to compensate for this difference, a higher strain is imposed 
upon the nickel steel, so that the final cost may not be increased. 
Accordingly, when it is used, the members are lighter, and the safety 
and durability of the structure are diminished. 

These remarks apply particularly to members in compression, 
where there is danger of column bending and secondary stress, but 
later it will be seen that they are not justified with regard to tension 
members. The writer is pleased to find that Mr. Waddell has arrived 
at the same conclusion by methods of deduction other than those 
which are about to be presented; and that, in order to realize economi- 
cal construction, he advises mixed structures, in which the compression 
members are of carbon steel and the tension members of nickel steel. 

The Banger Due to Column Flexure Stresses. — Previous to the 
investigations by Rankine and Considere, engineers did not compute 
the stresses due to column flexure; as a matter of fact, they did not 
possess sufficiently well demonstrated theoretical formulas to effect 
such computations. It was only by mere estimate, and by reliance 
on their personal experience, that they gave the members the shape 
and dimensions which appeared to them able to resist these stresses. 
But now that they have been startled by the number of great disasters 
which have occurred in various parts of the world, such as that of 
the Quebec Bridge over the St. Lawrence, the Tay Viaduct (England), 
the highway bridge at Morava (Russia), the temporary viaduct at 
Tarbes (France), etc., they are learning to be more careful, and every 
engineer, worthy of the name, computes with extreme exactness the 
stresses due to column flexure. In France, the designers of structures 
for the use of the public are positively compelled to do so by the 
Government, and a law, dated August 29th, 1891, provides that the 
maximum stresses for steel shall not exceed 16 500 lb. per sq. in. for 
main members, column flexure and wind effect included. 

In fact, even 16 500 lb. per sq. in. must be considered high for 
large structures. The writer, in his practice, has never dared to 
approach it in main members which have no intermediate supports. 

In truth, the assumptions involved in the deduction of the formulas 
are not always fully realized in practice. An infinite number of cir- 
cumstances may influence them unfavorably. In order not to extend 
this discussion uselessly, only two instances will be cited. 

It is assumed, for instance, that a member in compression is sup- 
ported normally and uniformly on its base. There is nothing perfect 
on earth, however, and, in a member of large cross-section, it always 
happens that an angle or some other point of the section carries a 
disproportionate strain, resulting in an unequal distribution of stress, 
which, again, induces flexure; and it is known that when buckling 
has begun, the loss of resistance of the member continues rapidly until 
final failure. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 369 

Moreover, in built-up members, like the chord which failed in the Mr. Amodin. 
Quebec Bridge, provision against column flexure is ordinarily made by 
using the radius of gyration of the entire cross-section, without con- 
sidering that each piece, of the several composing the member, owing 
to its small depth, is subject to the danger of individual buckling, 
which will result disastrously, step by step, in the buckling of the 
entire member. Everything points to the belief that these were the 
two initial causes of the Quebec disaster. It seems to the writer that 
it is chiefly in cases of this kind that nickel steel is out of place. 

English engineers, more than all others, discount the effect of these 
disadvantages, by allowing a low modulus of stress, and using very 
massive sections, following therein the traditions of R. Stephenson 
and Brunei. The English are great users of metal, and have learned, 
better than any other nation, how to produce it cheaply and in large 
quantities, and naturally are trained in the art of using it. 

On the other hand, French, German, and American engineers see 
the advantages of elegance and lightness. To enjoy these advantages, 
it is necessary to design with great precision, and leave nothing to 
approximation and the errors resulting therefrom. 

More than all others, American engineers have shown themselves 
to be daring in this tendency to extraordinary lightness. Many think 
that they have gone too far. The writer came to this conclusion when 
he was informed that the specifications for the monumental Blackwell's 
Island Bridge authorize, for nickel steel, a maximum stress of 30 000 lb. 
per sq. in. for the regular load, and 39 000 lb. for the congested load; 
and, for carbon steel, 20 000 lb. for the regular load, and 24 000 lb. for 
the congested load. He is of the opinion that such stresses are much 
too near the elastic limit of the material, in order to have the margin 
of safety which all public works should possess, and where the 
original provisions are so often exceeded by unforeseen and manifold 
causes, among which may be cited as the most frequent: 

pirst.— Increase in the moving load for which the structure was 
designed, made necessary by the peremptory demands of traffic, such, 
for instance, as compel railroad companies to have recourse to heavier 
and still heavier rolling stock. 

Second.— An erroneous valuation of the principal stresses due to 
(column) flexure. This was the case with the Quebec Bridge, where, 
as Professor Resal has shown, immediately before the collapse, the 
plates of the chord which failed must have reached a stress, which, 
including flexure, amounted to 28 500 lb. per sq. in.,* instead of the 
6 000 lb. which had been estimated for this point of the truss. In fact, 
Professor Eesal's computations would show a greater stress than 
28 500 lb. per sq. in., but he admitted the assumption of uniform 
distribution of stress over all the plates of the member, whereas it 
may be taken as cer tain that such a unifo rm^dis^ib ution cannot exist. 

• Encyolop6die des Travaux Publics, Vol. I, p. 582, Ch. B6ranger, Editor, 1908. 



370 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Arnodin. Third. — Increase of the fixed load, resulting either from errors in 
the original weight estimate or from strengthening the structure 
during the course of erection. This is the case with the Blackwell's 
Island Bridge, where the expert engineer investigators, Messrs. Boiler 
and Hodge, find that in certain members the calculated maximum 
stresses, with the weight of the structure as completed, reach 49 000 lb. 
per sq. in. for nickel steel instead of 39 000 lb., as originally specified, 
an increase of 26%; and for carbon steel, 35 500 lb. per sq, in., 
instead of 24 000 lb., an increase of 47 per cent. 

These examples, important in their significance, show that it will 
be wise to limit the maximum stresses in the metal, whatever its 
nature, to stresses much below those used in the United States, 
particularly for compression members in which the secondary stresses 
are difiicult to analyze and calculate with accuracy. 

Then, also, it is necessary to be so much more careful with com- 
pression stresses, since failure occurs as soon as the elastic limit is 
exceeded, whereas a tension member will continue its service long 
after that has taken place. 

Members in Tension.— \n members of this class the science of the 
engineer is simplified considerably by the fact that he needs to con- 
cern himself only with the main stresses, as secondary stresses and 
flexure are naturally eliminated by the tension itself. 

It follows, from the author's tests, that the most important property 
of nickel steel is its tensile resistance. Therefore he advises taking 
advantage of this quality by using the metal principally in tension 
members. 

This is equally true for carbon steel, and even for wrought iron. 
It is a fact which has already been demonstrated by a number of 
clear-sighted men. It is surprising, therefore, to note how little 
advantage engineers of all countries have taken of this chief quality 
of steel, that is to say, its tensile resistance; for, in nearly all metallic 
structures, it is the compression member which predominates. 

If it is necessary to show by facts the superiority of the tension 
member, it sufiices to give as an example the Brooklyn Suspension 
Bridge, 26 years old, which continues in service in spite of the fact 
that on many occasions its cables have had to carry stresses which 
exceed by far those for which they were designed. 

The point is equally well proven by the numerous photographs 
of the Quebec Bridge — after the failure — ^vhich show very clearly 
that the eye-bars of the top chord are the members which siiifered 
the least from the many strains occasioned by the fall, even though 
these very members were subjected to the greatest strains. A number 
of these eye-bars can be used again without any repairs. 

Lessons of Experience. — Tlie Quebec Bridge disaster, and also the 
errors of computation in the lilackwell's Island Bridge, furnish food 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 371 

for serious thought, and make it necessary that means be found Mr. Araodi 
to prevent, or at least to decrease the chance of, their recurrence. The 
engineer has a duty to perform — even as that of the physician is to 
check the progress of a fatal malady — and this is indicated by the 
fact that the investigating commission, consisting of Messrs. H. Hol- 
gate, J. G. G. Kerry, and John Galbraith, found that the errors made 
were not caused by insufficient technical education or by negligence, 
malice, or an excessive desire for economy. There is reason, then, 
to think that there is something wanting in the science of the 
engineer. 

This want does exist, perhaps to a greater extent in America 
than in Europe. The engineer devotes himself to his studies, and 
to extremely laborious computations, in order to valuate the stresses 
in the different members of the structure he builds; then, as soon as 
the structure is completed, he loses sight of it, and confides its mainte- 
nance and supervision to a subordinate who is incapable of observing 
whether the theoretical calculations are realized in practice. 

He thus resembles the captain of a vessel, who, before setting out 
on his voyage, carefully adjusts his compass, computes his bearings, 
and tests his steering apparatus, but who, as soon as the anchor is 
raised, leaves his ship, and entrusts its safety to an ordinary sailor, 
who, perhaps, has been trained to keep on a course as indicated by 
the compass, but who would be entirely at a loss as soon as unforeseen 
circumstances, such as wind, currents, etc., set in to divert the ship 
from its course. Such a captain would be considered guilty of negli- 
gence. Is the engineer entirely free from this reproach? 

The French Government was the first to take up this matter. 
Again and again it has called upon its engineers to make frequent 
inspection tours to all the public works of their districts. It even 
orders them to revise their theoretical computations, so as to take 
account of the changes which have been made in the structure itself, 
or in the traffic upon it, since its completion. In other words, it 
compels the captain to make his observations at stated intervals, in 
order to verify his position. In this way the number of possible 
disasters is limited. 

Is that sufficient? The writer does not think so; at least, not for 
structures of such great size as the Forth Bridge, the Quebec Bridge, 
the BlackwelTs Island Bridge, and a number of others, principally in 
the United States. For, if there should be any error in computation 
or any unnoticed movement of the foundations under the supports, or 
any error in the assumptions, the safety of the bridge may become 
impaired, unless some visible sign gives warning. 

In the writer's opinion, a large, well-planned metallic structure 
should be provided with an "indicator" for showing the stresses, just as 
a well set up steam generator has a gauge to indicate the pressure. 



372 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Arnodin. or as an electric conduit must be provided at certain points with volt 
meters to give the intensity of the current, and ampere meters to 
show its quantity, even if the boiler or the electric conduit possesses 
all the desired capacity for the service imposed on them. 

These opinions may possibly clash with adopted ideas, or routine, 
but that is the fate of all new things. However, this ought not to 
stay the engineer, who, by his very profession, is a man of progress. 

It remains to demonstrate that these ideas may be realized. 

The writer shows here, for the first time, that, in order to study 
his structure, the engineer has only to introduce a "flexible member" 
Ipiece soupW] at some proper point of the bridge. This flexible 
member will act as an indicator to show, by indisputable external 
signs, what actually occurs in the structure at a given point, and, 
consequently, will enable the engineer to deduce, either by simple 
composition of forces, or by a graphic diagram, what takes place in the 
adjoining members. 

It should be noted here, that this "flexible member" can only be 
placed where the stress is tensile. However, excepting the arch, which 
is in compression only, bridges of all classes contain tension members. 

The foregoing discussion shows that the resistance to tension is 
the most important property of steel. By taking advantage of this 
property, in planning a great work, a technical and economical benefit 
may be realized, while, otherwise, work is done under unfavorable 
conditions. 

The Flexible Member. — In order to make the action easily under- 
stood, take any construction, such as a derrick. Fig. 78, for example, 
composed of a vertical post, DC; an inclined boom, A C; an inclined 
tie, E D; a vertical hanger, A B, which carries an unknown load, B; 
and, finally, a flexible member, A D, acted upon only by a tension 
stress denoted by t^ ; the whole forming an articulated system lying in 
the same plane. 



Fig. 78. 



After t^ has been determined, it is known that the tension, t, in 



A B is 

the tension in D .E^ is 



a 



t. = t.Jl+~ 



DISCUSSION ON NICKEL STEEL TOR BRIDGES 



373 



the compression, Cj^, in A C becomes 



Mr. Arnodin. 



- = ^n1 



1 + 



h^ 



the compression, c^, in D C becomes 



c , = ^ 



The stresses may also be found graphically, as shown in Fig. 79. 







FiQ. 79. 

The problem is now to find the value of t-^. 

It was to solve this problem, while making experiments for the 
construction of the ferry bridge at Rouen, that the writer conceived 
the "Tension Meter," the principle of which, under the name of "Test 
Cable," he explained before the Societe des Ingenieurs Civils de 
France,* which principle is herewith reviewed. 

The Tension ilfeier.— Since t^ (Fig. 78) is in tension, a flexible 
cable may be used in its place; and it is perfectly reasonable to use it 
because, of all material at the disposal of the engineer, the cable, con- 
sisting of (nickel or carbon) steel wires, offers, with a minimum of 
weight, a maximum of resistance and safety. Therefore, this system 
involves a structural advantage rather than a disadvantage. 

As the coefficient of elasticity of the steel in cables is very nearly 
the same as that in eye-bars, and as the variation in the curve of the 
cable within the limits of the variation of the load affects the total 
elongation very slightly, the increase in the flexibility of the structure 
will in no case result in a serious disadvantage. 

t^ being a cable, will assume, within its free length, the curve of a 
catenary under the twofold effect of its own weight and its tension. 
No human jwwer and no secondary stress can oppose this natural law, 
which acts vipon all the particles of the flexible member. 

Suppose a longitudinal, and extremely thin, element of the cable 

to be separate from the entire free length of the member. To fix the 

idea conceive this elementary wire to have a cross-section of 1 sq. mm. 

The density, the flexure, and the tension of this wire, per unit section, 

*At the meeting of February 20th, 1903. 



374 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



Mr. Arnodin. are the same as those of the whole member. Therefore, it is only 
necessary to apply a dynamometer to this wire in order to observe the 
strain on the member; that is to say, it will show the actual stress, 
which includes not only the main stresses, but also any non-analyzable 
secondary stresses which may occur. In short, this elementary wire, 
conceived as separated from the remainder of the member, will not be 
the indicator of any theoretical deductions, but will show directly and 
accurately the actual condition of the member. 

However, in practice, one cannot realize the conception of the 
isolated cable element, hence this difficulty must now be overcome. 

This is accomplished simply and easily by adding, at the side of the 
flexible member, a very thin wire. 

If this wire is of the same specific gravity, and if its free curve is 
the same as that of the member, then its tension must necessarily be 
that of the whole member. The amount of this tension, indicated by a 
dynamometer, will certainly throw light on the actual stresses. 

Strictly speaking, the addition of the tension meter wire decreases 

slightly the stress of the member, since it takes its portion of the total 

stress; however, the error is a negligible quantity, and only has a value 

s . . . 

equal to — ^, in which 8 is the total section of the flexible member, and 

S 

s the section of the wire of the tension meter. 

In practice, the Arnodin tension meter consists of a wire, the 
cross-section of which is 1 sq. mm. It is of the same specific gravity 
as the main cable member, and at one end there is a light dyna- 
mometer, which in turn is attached to a regulating screw, with which 
the wire is adjusted to the same curve as the main cable in its unsup- 
X)orted length. 




Fig. 80. 

It should be noted that the greater the unsupported length of the 
cable, the greater is its curve, and the easier and more precise will be 
the observations. The flexible member, therefore, should be of great 
length, and for tliis reason bridges of large span are more favorable for 
this installation than those of smaller span. In sucli bridges, also, the 
obsen^ations are most valuable. 

Further, the flexible member should always be either horizontal or 
diagonal, never vertical, for the method is based on the curve of the 
cable, which does not occur in vertical tension members. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 



375 



Suspension bridges carried by cables of flexible wires, are, by their Mr. Arnodln. 
very nature, admirably adapted to this kind of testing, for in such 
bridges there can always be found a free inclined length of cable — the 
anchor span, for example — in which the tension meter may be placed; 
and it will be difficult to understand, once the system has become 
known, how engineers who are responsible for these great structures 
can afford to deprive themselves of this experimental control over the 
stresses actually carried by the cables. 

In cantilever bridges. Fig. 81, such as the Forth, the Quebec, the 
Blackwell's Island, etc., nothing is simpler than to make of cables 
either of the members, A B or A C, which, being always in tension. 




Fig. 81. 

can be used as the "flexible member," so that its stress may be tested 
by the tension meter, both during construction, and later, during 
maintenance; and, from the observations, the stresses in the other 
membei*s may be deduced. 

It may be stated that if this precaution had been taken, there 
would not have been cause to deplore the Quebec disaster, nor the 
embarrassing situation of the Blackwell's Island Bridge; for, during 
the course of erection, the indications of the tension meter would have 
given warning of the errors in the computations, and in the assumptions 
upon which the design was based. 

But even when a structure has been designed and erected with the 
greatest possible care, it would still be of benefit to be able to make 
sure whether or not it behaves according to the design, and whether, 
subsequently, dynamic strains or secondary effects do not modify the 
computed stresses. 

In the writer's opinion, it would be very satisfying to be able, by 
the use of the tension meter and its gauge, to control the stresses of 
the Brooklyn Bridge when the traffic is at its maximum, or in the 
Forth Bridge, under a passing train, or in any other bridge of 
lesser span. 

The French engineer, M. Leinekugel Le Cocq, who is a specialist 
in work of this class, has mentioned the remarkable results which he 
obtained by the aid of the tension meter, while in charge of the testing 
of the ferry bridges at Marseilles and at Brest,* which results he 
wisely compared with the results of the theoretical computations. 
* Le Ghiie Civil, February 34th, and March 4th, 1906, and October 31st, 1908. 



376 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Arnodin. It Ought to be realized that the extension of this system to all 
large structures would greatly promote that progress which engineers 
desire, for it would indicate by exact methods where the computations, 
or the theoretical hypotheses, were at fault, would lead to their rectifi- 
cation and would also clear up some disputed questions. 

It would insure for the future a higher degree of safety in the 
execution of structures, a more judicious use of material, a better 
economy, and, above all, a greater confidence and sense of security 
on the part of the public. 

Rigid Compression or Tension Members. — ^Many engineers have 
been on the lookout for a method of observing the actual strains, but 
not having at their disposal the "flexible member" and its measuring 
apparatus, because it was not known to them, they have attempted to 
observe the strain on rigid compression or tension members. In such 
a case one encounters a difficulty in principle, namely, the fact that the 
manifestation of strain is made possible only by the elastic property 
of the material, which is generally assumed to take a strain of ^^-o^ny 
of its length for every kilogramme per square millimeter (yt-uvvtovs 
per pound per square inch), tension or compression. 

It follows that every error of observation, and every secondary 
consideration which affects the measuring apparatus will be multiplied 
by 20 000, and this might often lead to grave mistakes. 

Again, the accepted value of ^q-oott P^r kilogramme for the elastic 
elongation is not uniform. It varies with the temperature of rolling, 
with the chemical composition of the steel, and with many other un- 
avoidable causes. It may thus happen that, of the many members 
in a bridge, those on which the tests are made possess an elastic co- 
efficient other than the usual value. 

Moreover, this method will only give the stress due to the load 
variation, and not the total stress. Thus it may happen that the 
member under observation is already strained so highly that it ought 
not to receive an additional stress due to an increase of load, and this 
additional stress is the only one which can be measured. By this 
method of observation, no real assurance of the degree of safety of 
the structure is obtained. 

Very ingenious apparatus have been invented for such measure- 
ments. The best of these are the "Manet", and the gauges by Professor 
Rabut, of France, which make it possible, according to the picturesque 
words of the inventor, to "enquire of the bridge how it is getting on." 

It follows, from the foregoing, that the engineer should not lose 
sight of his bridge as soon as he has completed its construction, but, 
like the good teacher who follows the life of his pupil, he ought to 
keep track of its career, observe its qualities and its defects, and try 
to remedy the latter, or at least deduce therefrom lessons which will 
help him in future structures. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 377 

As to "Removability." — The science of conserving the life of the Mr. Arnodin. 
stnictiire has not been exhausted in the preceding remarks. 

In the case of small bridges, involving a low cost of reconstruction, 
durability is perhaps not so important a consideration, for when the 
life of the structure is at an end, it is relatively a simple matter to 
erect in its place a new bridge. 

But, in the case of structures of such great spans as have been 
considered herein — and in the future the spans may be greater still — 
which absorb millions of the public funds, it is necessary to be more 
far-sighted, and to provide at the very outset for a much longer life 
than is ordinarily assigned to metallic structures. 

Up to the present there is not much definite information regarding 
the durability which a metal structure in the free air ought to possess. 
It is principally dependent on the degree of care in maintenance, 
and many of the best engineers believe that metallic bridges as now 
constructed will not endure for one hundred years; in fact, there are 
a number of bridges which should be renewed after having reached 
the age of thirty years. 

This fact led the writer to introduce into suspension bridges, the 
principle of "Removability" {I'amovihilite], that is, to arrange the 
structure, at the outset, so as to be able in the future to replace the 
several parts, without affecting unfavorably the resistance of the whole, 
and without interrupting even temporarily the regular traffic on 
the bridge. 

Accordingly, the structure may be made to last indefinitely by re- 
placing individually, one at a time, as may be required, each of its 
members, just as a railroad track can be made to exist as long as the 
ground that supports it, by continually renewing the worn-out rails. 

The principle of "Removability" is not yet applicable to all classes 
of bridges. It is particularly adapted for tension members, and, like 
the latter, for the "flexible member." 

Danger of Inaccessible Parts. — The numerous examinations of 
old bridges made by the writer at the beginning of his practice, have 
led him to the conclusion that the greater part of the failures due to 
deterioration resiilt from parts inaccessible to inspection or mainte- 
nance. Accordingly, he has spent thirty years of his career in making 
war upon the inaccessible parts of bridges, notably upon anchorages 
of suspension bridges, which, being generally in low and damp places, 
are most exposed to deterioration. 

It is necessary, therefore, that the members of the anchorage should 
be entirely accessible, so that the engineer, by frequent inspection and 
by proper maintenance, may be enabled to repair the deterioration in 
time, and to see the trouble before a disaster unexpectedly reveals it. 

Conclusions. — From the preceding considerations the writer reaches 
the following conclusions: 



378 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Arnodiu. First. — That the use of nickel steel may be advantageously recom- 
mended for the tension members of bridges. 

Second. — That its superiority over carbon steel should be utilized 
for the increase of safety rather than for the diminution of section. 

Third. — It appears that engineers in the United States have been 
too daring in their specifications for allowable maximum stresses. 

Fourth. — That it is advantageous, in the study of design, to devise 
forms of articulation vphich will utilize the tensile resistance of the 
metal, because tension eliminates flexure and a number of secondary 
stresses. 

Fifth. — ^That the arrangement of certain tension members should 
be such as to form flexible members, which, by the aid of the tension 
meter, will enable the precise observation of the actual stresses. 

Sixth.^ — It would be advantageous to have a bureau of control to 
test metallic bridges in order to make sure that the specified stresses 
are not exceeded, and that the deterioration of the metal does not 
affect their safety. 

Seventh. — That no part of a bridge should be inaccessible to inspec- 
tion and maintenance. 

Eighth. — That those principal members, upon which the safety of 
the bridge depends, should, as far as possible, possess the advantage of 
"removability." 

Regard for these conditions will add to the safety of bridges, and 
will greatly increase the confidence in the science of the engineer, 
and in the structures which he executes, no matter how great the 
apparent boldness of their design may be, 
Mr. Worsdeil. WiLSON WoRSDELL, EsQ.* (by letter). — The following particulars, 
relating to a nickel-steel fire-box which is being built for the North 
Eastern Railway, may be of some interest. The composition of this 
steel is as follows : 

Nickel .' 16.3 per cent. 

Carbon 0.54 " 

Manganese 2.87 " 

Phosphorus 0.027 " 

Sulphur 0.032 " 

Silicon 0.378 

This steel gave a maximum ultimate stress of 40 tons per sq. in., 
with an elongation of 50% in 3 in., whereas an ordinary mild-steel 
sample of similar dimensions gave 29.6 tons per sq. in., with an 
elongation of 35% in 3 in. A strip of the nickel steel, ^ in. thick, 
was bent double, cold, without cracking. When flanging the fire-box, 
the plates were annealed by heating them to redness and plunging them 
in water, the carbon having no hardening effect, due to the presence 
of the nickel. 

* Chf. Mech. Engr., North Eastern Ry., Great Britain. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 379 

Although the steel showed such ductility under test, it was most Mr. Worsdeli. 
difficult to machine and press. This was probably due to the 
manganese. 

The writer was under the impression that the addition of nickel 
did not make steel any more difficult to machine, and that this had been 
demonstrated by Iladfiold, although some American manufacturers 
consider that the nickel is responsible for this difficulty. 

WiLLTAM F. Pettigrew, Esq, (by letter). — The writer, having charge Mr. Pettigrew. 
of the maintenance of several bridges and viaducts, has often thought 
that nickel steel could be used for such structures, particularly with 
reference to decreasing the corrosion between wind and water, that 
is, where the material is at times covered with water and then exposed 
to the atmosphere. 

In two of these viaducts the columns are of cast iron braced by 
wrought-iron tie-rods; in one case 1^-in. and, in the other, 2-in. rods 
are used. A space of 7^ ft., embracing- parts of these rods, is exposed 
to water and the atmosphere intermittently, and at certain places 
there is corrosion. The writer has often thought that, if these tie-rods 
were of nickel steel, this difficulty of corrosion would be considerably 
overcome, but the trouble has been to obtain nickel-steel bars, although 
ai)plication has been made to many manufacturers. This is confirmed 
by Dr. Waddell, as he states that up to the present time the only 
nickel-steel bars manufactured, within his knowledge, have been made 
by the American Bridge Company. The results he has obtained in his 
tests are certainly most satisfactory, and will be a great boon to 
engineers in the future. 

In reference to the viaducts under the writer's control, the corrosion 
and salt tests are very interesting, and prove that in both cases the 
loss with nickel steel is much less than with ordinary carbon steel. 

J. A. L. Waddell, M. Am. See. C. E. (by letter), — In beginning Mr. AVaddeii. 
this resume the writer desires to tender his hearty thanks to all who 
have been so courteous as to discuss his paper. It is true that he had 
hoped for a much larger discussion, especially from certain prominent 
American bridge engineers whose opinions on the various matters 
treated in the paper would certainly carry great weight; but he 
recognizes that the paper is difficult to discuss, since it treats of a 
subject which is almost entirely now in engineering. The writer feels 
especially indebted to the various foreign engineers who have done 
him the honor to discuss the paper, and begs to tender them herewith 
the assurance of his deep appreciation of their consideration and 
courtesy. 

In some of the discussions certain side issues of much value are 
treated; but the writer deems it best, in order not to extend the dis- 
cussion unnecessarily, to omit practically all reference to these in 
his resume. 



380 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Waddell. Again, he takes cognizance herein of only one-third of the discus- 
sions. By so doing he does not by any means intend to imply that the 
others are unworthy of notice, but brevity demands that he refer only 
to those points upon which he differs materially from the various 
writers. Their discussions will be treated hereinafter in alphabetical 
order. 

In bridge designing there are certain restrictions governing mini- 
mum sections and minimum thicknesses of metal, which all thoroughly- 
written bridge specifications recognize, consequently Mr. Arnodin's 
criticism of nickel-steel bridges on account of possible ultra-small 
sections and ultra-thinness of webs will not hold. Moreover, in com- 
puting the weights of metal and the comparative costs of spans given 
in the diagrams, the writer took due account of these restrictions in 
designing. 

Mr. Arnodin's objection to nickel steel because of its tendency to 
increase vibration due to reduction of total weight of structure will 
apply properly to short, light spans, such as those of county bridges; 
but in modern railway bridge designing the live loads are so great 
that the main members and connecting details of spans of even ordi- 
nary length become bulky and clumsy. The effect of the use of nickel 
steel would be to reduce this objectionable feature of modern bridge 
designing. 

Mr. Arnodin has a wrong notion in mind when he states that the 
use of nickel steel will reduce the safety and durability of a bridge. 
The writer's whole economic investigation is based upon keeping the 
safety, strength, and durability of both carbon-steel and nickel-steel 
bridges alike, and ascertaining the relative costs of the two kinds of 
superstructure. The specifications of the paper were intended to be 
drawn so as to keep these attributes as nearly identical as practicable, 
and the writer believes that they will accomplish that purpose, except, 
perhaps, that the durability of the nickel-steel structures will be 
greater than that of the carbon-steel structures, because of the alloy's 
superior resistance to corrosion. 

Mr. Arnodin is in error when he states that the writer advises 
"mixed structures in which the compression members are to be of 
carbon steel and the tension members of nickel steel," for in his 
"mixed metal" bridges he suggests using carbon steel only in those 
parts where the adoption of the stronger metal would effect no good 
purpose or economy. Nickel stool is just as fit for compression mem- 
bers as carbon steel, provided proper consideration be given to the 
designing and detailing. Compression members which either fail or 
do not develop the proper strength are deficient because they lack 
proper design and detailing, and not because they are compression 
members per se. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 381 

Mr, Arnodin's statement that the stresses allowed on nickel steel Mr.Waddell. 
in the Blaekwell's Island Bridge are too high is, in the writer's judg- 
ment, correct; because the greatest intensity for live load, impact 
allowance load, and dead load should not have exceeded 30 000 lb. on 
eye-bars, while 39 000 lb. was adopted. Even the impossible combina- 
tion of greatest assumed live load, impact allowance, dead load, and 
wind load should not stress the eye-bars higher than 37 500 lb. 
per sq. in. 

It would not be scientific engineering to lower the intensities of 
working stresses in a bridge in anticipation of a possible augmenta- 
tion of dead load during manufacture and construction, because every 
bridge should be designed complete in every detail before the drawings 
are sent to the shops, and it is the duty of every bridge designer to 
check the dead load of every structure from the completed drawings 
before letting the latter pass out of his possession and control. 

It is not the writer's intention to discuss Mr. Arnodin's "Tension 
Meter," but he cannot see how such an apparatus placed at the top of 
the tower in the Quebec cantilever bridge could have indicated an 
impending failure in one of the bottom chords. 

The writer does not agree with Mr. Arnodin's suggestion that 
bridges should be designed so that all their parts may be removed and 
replaced as they wear out, for this would involve great lack of 
rigidity as well as a multitude of constructive complications; but he 
would prefer to utilize the principle adopted by the designer of the 
famous structure known as the "Deacon's One-Horse Shay," in which 
all parts were equally strong, gave good service without repairs for 
many years, and then suddenly went to pieces because of the uniform 
deterioration and simultaneous collapse of all its members. 

Mr. Carpenter's dread of mixing the steels improperly, in a bridge 
built of both nickel and carbon steels, is unnecessary. It was not the 
writer, but his inspector, Mr. Saunders, who on page 248 said that 
"extreme precautions were taken at every step to keep the two steels 
separate." These precautions were adopted by the inspector so as not 
to spoil the results of the tests of apparently identical columns made 
of the two kinds of steel. In any well-organized shop the marking 
system will make it certain that the right kind of steel is used in each 
place in mixed steel bridges. 

If one were dealing with the usual English ratios of depth of truss 
to length of span, viz., about 1 to 12, instead of the customary 
American ratios of about 1 to 6, or 1 to 7, he might have cause to 
worry over the increase in deflection due to using nickel steel ; but the 
deflections of modern American bridges are so slight that this matter 
is hardly worthy of consideration, in so far as they affect the camber 
and the track. 



382 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Waddell. There is something in Mr. Carpenter's statement that the greater 
deflections of nickel-steel structures will augment the secondary stresses 
as compared with the equivalent carbon-steel structures; but, on the 
other hand, the smaller dimensions of the nickel-steel members will 
have a contrary effect. Much has yet to be learned about the gravity 
and extent of secondary stresses in bridges, and how best to avoid or 
provide for them. 

Mr. Carpenter evidently rivets his stringers together continuously 
from end to end of span, even in long-span structures ; but the writer's 
practice is to insert occasional expansion pockets in his floor-systems 
so as to prevent the stringers from doing some of the work of the 
truss chords. 

As for the greater distortion of nickel steel affecting the adherence 
of the paint — does not such a thought involve the stretching of one's 
imaginative faculty beyond the elastic limit? 

If there is any other alloy of steel that will give as good results 
as nickel steel at less cost, the Profession ought to know it. The 
writer did look into vanadium steel, but discovered that it is altogether 
too expensive for bridgework. If vanadium could be furnished at any 
reasonable cost, vanadium steel might compete successfully with nickel 
steel. 

As for the accuracy of the Phoenix testing machine, the writer, 
when writing the paper assumed that the numerous testing machines 
used in the investigation by his various inspectors gave correct records. 
It would, of course, be desirable to have this doubt about the Pha3nix 
records of column resistances removed; but, unfortunately, the writer's 
time is so taken up with important professional work that it is im- 
practicable for him to make the investigation, especially as the time 
limit set for the completion of this resume of discussions is nearly 
reached. 

In any case the variation of the machine wovild not affect very 
greatly the ratio of strengths of nickel-steel and carbon-steel struts, 
even if the actual strengths recorded were wrong; and it must be 
remembered that the specifications for nickel-steel bridges and the com- 
parisons of cost made by their use were based on the comparative 
strengths of nickel steel and carbon steel. 

It seems to be a slur upon the Engineering Profession that there 
should be any doubt about the correctness of results from such a large 
and important testing machine as the one at Phoenixville; and it is 
to be hoped that when the great machine upon which Mr. Emory is 
now working is completed, engineers will have at their disposal an 
apparatus which may be relied upon absolutely. 



DISCUSSION ON NICKEL STEEL FOR BRIDGES 383 

The writer investigated the use of ferruginous nickel, mentioned Mr. Waddell. 
by J\Ir. Fowler, but found that, unfortunately, it contains so much 
copper as to prohibit its use for manufacturing nickel steel without 
first separating the nickel. 

The writer does not believe that the acid open-hearth process will 
produce any better high-grade steel for bridges than the basic open- 
hearth process as now operated; in fact, in his opinion, the latter is 
preferable. 

Mr. riallsted's criticism of the increase in rivet diameter not being 
applicable to the flanges of channels is correct; but the writer would 
state that, unless American manufacturers decide to roll channels 
deeper than 15 in., there will be but little use for these sections in 
nickel steel. For some years, at least, nickel steel for bridge building 
will be confined to long spans, where channels of obtainable dimensions 
are inadmissible. 

In plotting his economic curves, the writer did not forget the 
decrease in economy due to the increased sectional areas made neces- 
sary by the larger rivet holes in nickel steel. 

Mr. Hallstcd's various anticipations of increased cost are mainly 
those which will exist in the transition stage, while carbon steel is 
being gradually supplanted by nickel steel. They would soon be for- 
gotten, as were the similar anticipated troubles when carbon steel was 
about to replace wrought iron for bridgework. 

Mr. Hallsted has made an error in quoting the writer's "De 

Pontibus" formula for top chords. He assumes it to be 16 000 — 70 

r 

and compares that with the nickel-steel formula of 30 000 — 120 , 

r 

while it should have been 18 000 — 70 . This explains why he thought 

he found the writer in error. 

Mr. Lindenthal is mistaken when he says that a large demand for 
nickel would probably raise its price. On the contrary, such a demand 
would lower the price materially, because the supply of nickel ore 
already in sight is immense, and an increased demand would certainly 
result in improved methods of extracting the metal from the ore. 

The writer does not agree with Mr. Lindenthal when he says that 
the use of high steels in bridge construction is justified only in very 
long spans, because (as previously stated) modern live loads have 
become so great that in spans of moderate length, especially for rail- 
road bridges carrying more than one track, the main members are 
becoming too bulky and the connecting details in riveted structures 
too clumsy. 



384 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. "Waddell. The deflections of modern, riveted bridges which are correctly 
designed, and in which the camber is halved by proper adjustment of 
ties, are of such minor importance (excepting the effect of secondary 
stresses) that they may be forgotten, whether the material of the super- 
structure be carbon steel or nickel steel. 

The writer agrees with Mr. Lindenthal that "to use high steel 
[including nickel steel] in a light structure, for the sake of low first 
cost, is not true economy," and it was on this account that he sug- 
gested that the alloy should not be adopted for highway bridges, unless 
these be of an unusually heavy character. 

The writer does not agree with Mr. Perry that experiments should 
be made upon the effect of continued vibration on nickel steel, because 
it is generally conceded to-day by the leading bridge engineers that 
the effect of repeated stress on bridge members, provided the elastic 
limit is not exceeded, is absolutely nil. The idea of deterioration of 
steel bridges by vibration or by fatigue from repetition of stress is a 
bugbear that has, fortunately, been relegated to the past, and the 
oppressive nightmare which it involved no longer disturbs the slumbers 
of the bridge engineer. 

In speaking of the cost of fabricating nickel steel, Mr. Prichard 
states that the writer's estimate is too favorable to the alloy. The 
writer, in proof of the correctness of his figvires, begs to quote the 
following letters from two high authorities upon such questions : 

"September 3, 1908. 
"Mr. J. A. L. Waddell, 

"608-11 New Nelson Bldg., 

"Kansas City, Mo. 

"Dear Sir: Referring to your letter of the 25th ultimo relative 
to our experience with nickel steel for bridge work, we beg to state 
that in so far as the ordinary operations of ship building are concerned 
we find that the cost of working nickel steel compares very favorably 
with that of mild steel. Our experiments were not made in such 
detail as to make a discussion by us profitable. 

"We have, however, worked nickel steel in government vessels for 
a considerable period, and have no difficulty what^^ver with it in the 
usual shop manipulations. The cost of drilling is slightly greater than 
that of medium steel. 

"We feel confident that the material you propose may be used in 
bridge work to great advantage in lessening the weight without 
materially increasing the cost. 

"Very truly yours, 
"The William Cramp & Sons Ship & Engine Building Company, 

"w. a. dobson, 

"Naval Architect." 



DISCUSSION ON NICKKL STEKL FOR BlflDGKS 385 

"Pittsburg, Pa., Dec. 26tii, 1907. Mr. watuieii. 
"Waddef.l & Harrington, 

"Mr. John Lyle Hjvrrington, 

"Kansas City, Mo. 

"My dear Mr. Harrington : Some time ago you wrote us iu regard 
to the increased cost of the shop work on nickel steel over carbon steel. 

"Since receipt of your letter we have made some experiments, and 
as near as we can tell from the experiments made, the additional cost 
of shop work will be about 10% higher for nickel steel than for 
ordinary steel. 

"We will be pleased to receive your opinion as to the probable 
difference in cost in working this material. 

"Yours truly, 
"McClintic-Marshall Construction Co., 

"H. H. McClintic, 
"7. P. & Gen. Mgr." 

Mr. Prichard, like some of the other engineers who have discussed 
the jjaper, fears that trouble may be encountered in the shops by the 
difficulty in distinguishing between nickel steel and carbon steel. 
This anticipation is needless, for the instant any tooling or shop 
manipulation of any kind is started the nickel steel will declare itself; 
moreover, the exterior surface of the alloy can generally be relied on 
to indicate its character. 

Mr. Prichard also fears that, in the future, inspecting engineers 
will not be able to determine whether a bridge was built of carbon 
steel or nickel steel. Again, the anticipation is unwarranted, for a few 
applications of a file would tell the tale; moreover, all railroad com- 
panies now-a-days keep office records of their structures in which the 
characteristics of the metal are given, as do also the manufacturers of 
such structures. In the examinations of old railroad bridges which the 
writer is constantly making, he experiences very little difficulty in 
ascertaining what structures or parts of structures were built of steel 
and what of wrought iron. 

Mr. Prichard states that: 

"Members in compression, in addition to resisting the effort of the 
load to crush them, have to resist its tendency to buckle and wrinkle 
them, and the resistance to these tendencies is about the same for steel 
of all grades." 

Had Mr. Prichard read carefully the writer's description of his 
experiments on full-sized columns, he would not have made such a 
mis-statement. Nickel steel resists "buckling and wrinkling" in 
columns, up to working limits of length, far better than carbon steel. 

The writer confesses that his specifications for nickel-steel eye-bars 
are not based upon complete experiments, and he regrets deeply that 
he was unable to procure proper eye-bar steel for testing; nevertheless, 
he feels confident that, when a comprehensive, scientific study is made 



o86 DISCUSSION ON NICKEL STEEL FOR BRIDGES 

Mr. Wadfieii. of the best composition and method of manufacture for nickel-steel 
eye-bars, it will be found that his specifications for these members will 
not be far astray. 

About the first investigations that would have to be made before 
using nickel steel in any great bridge are the composition of eye-bar 
steel and the proper method of annealing nickel-steel eye-bars. There 
is yet nuich to be learned about both these matters, and it is. to be 
hoped that the engineers of the Quebec Bridge will soon inaugurate 
such investigations. 

Mr. Prichard terms the unit stresses of the "De Pontibus" specifica- 
tions "high." He seems to forget that they are for equivalent static 
loads. As such, they are not high, compared with the general practice 
of the leading American designers and manufacturers of medium-steel 
bridges. 

The reasons why the writer advocates the basic open-hearth steel 
for the manufacture of nickel steel for bridges are: 

First. — This process is used almost exclusively in the United States. 

Second. — The basic process permits of cutting down the percentage 
of phospliorus to an exceedingly small amount, ■while the acid process 
does not; and experience has shown that phosphorus is especially objec- 
tionable in nickel steel. 

Mr. Ross is right in stating that the corrosion experiments should 
be carried further. The writer would suggest subjecting both nickel 
steel and carbon steel to the deteriorating influences of the salt air 
and salt water of the Gulf of Mexico, where he has seen railway rails 
corrode in a few years to such an extent as to be almost unrecognizable 
as rails. 

The writer said nothing about the use of nickel steel for reinforcing 
concrete because he does not believe it is well fitted for the purpose. 
Ordinary medium steel is strong enough in all conscience — and pos- 
sibly too strong, considering the adhesive strength of the mortar to 
the metal. The use of nickel steel would reduce the ratio of strength 
of adhesion to strength of metal, and would thus intensify the most 
serious reason for apprehending that the life of reinforced concrete 
structures is limited. 

In reference to the use of nickel-steel rivets in nickel-steel plates 
and carbon-steel rivets in carbon-steel plates, to which Mr. Sparrow 
calls attention, the writer would state that his remark applied mainly 
to test specimens; for it is obvious that in a bridge no harm could be 
done by using nickel-steel rivets to connect carbon-steel stay-plates or 
lacing bars to nickel-steel main members; but, in lateral struts, having 
both the sections and the details built entirely of carbon steel, it would, 
of course, be preferable to use carbon-steel rivets. 



DISCUSSION ON NICKEL STKEL FOR BKIDGES 387 

Til (•(•iicliisioii, there are certain points to which the writer desires Mr. Waddell. 
to call attention. 

In the paper as printed herein much more attention is given to 
mixed-steel bridges than to bridges built entirely of nickel steel; but 
in the original paper, which is on file in the Society's Library, both 
classes of structure receive the same consideration. When it became 
necessary, on account of economy of space and cost, to reduce the 
original paper to much smaller dimensions, the writer and the Coin- 
niiltee to whom the pajjcr was referred concluded that, as the transi- 
tion from carbon-steel bridges to nickel-steel bridges would probably 
take place through the medium of "mixed-steel" bridges, it would be 
best to omit all the comparative cost diagrams relating to bridges 
built of nickel steel throughout. It is possible that in some future 
publication the writer will print these omitted diagrams; for, if nickel 
steel ever does begin to replace carbon steel for bridgework, the time 
will certainly come when for various good reasons carbon steel will not 
be mixed with nickel steel in the same span. 

The next step to take in order to determine the desirability of the 
use of nickel steel for bridges is for some engineer, who has a large 
bridge to build, to make preliminary plans and specifications for super- 
structures of both nickel steel and carbon steel and call for bids thereon, 
thus ascertaining with certainty the present economics of the two 
metals; then, if nickel steel shows a great advantage in price, it will 
be necessary to make some further tests, mainly of eye-bars and full- 
sized columns, so as to settle finally the proper intensities of working 
stresses for such members in advance of the preparation of the work- 
ing plans. 

The adoption of nickel steel is really being forced upon the Profes- 
sion by the call for bridges of exceedingly long span to support large 
live loads. In the stiffening trusses of the great Manhattan Bridge 
it proved necessary to adopt nickel steel to keep the weight of the 
structure down to reasonable limits; and, in the writer's opinion, such 
a course will be found obligatory in the reconstruction of the long 
cantilever bridge at Quebec. The main span of this structure, 1 800 ft., 
is very close to the greatest practicable limit for cantilever bridges of 
carbon steel. The writer's calculations show that, at present prices 
of carbon steel and nickel steel, a saving of from 25 to 30% in the 
cost of the superstructure could be effected by adopting nickel steel in 
those parts where its use would be advantageous or economic. Conse- 
quently, the committee of engineers in charge of the work will no 
doubt carefully consider the advantages of nickel steel before deciding 
filially upon the character of the new s\iperstructure. 



AMERICAN SOCIETY OF CIVIL ENGINEEES 

INSTITUTED 1852 



TRANSACTIONS 



Paper No. 1104 

THE IMPROVEMENT OF THE OHIO RIVER. 

By William L. Sibert, M. Am. Soc. C. E.f 



With Discussion by Messrs. Tiieuun M. Ripley, ani> 
William L. Sibert. 



The recent completion of Locks and Dams Nos. 2, 3, 4, and 5, Ohio 
River, creating, in connection with Locks Nos. 1 and 6, a navigable 
depth of 9 ft. from Pittsburg to Beaver, Pa., a distance of about 30 
miles, and constituting the first completed section of the 9-ft. project, 
caused the writer to think that a review of the various engineering 
projects proposed for the improvement of this river, might be of 
interest to the Profession. 

The writer had charge of the construction of Locks Nos. 2, 3, 4, 5, 
and 6, from 1903 to the spring of 1907. Captain E. N. Johnston, Corps 
of Engineers, U. S. A., one of his assistants at that time, aided 
materially in preparing this paper. 

Physical Characteristics. 

The Ohio River is formed, at Pittsburg, by the junction of the 
Allegheny and the Monongahela Rivers. Its total length, all of which 
is navigable, is 967 miles. The navigable length of its tributaries is 
about 3 000 miles. The drainage area above Pittsburg, including the 
valleys of the Allegheny and the Monongahela, is about 19 000 sq. 

* Presented at the meeting of December idtli. 1!«>S. 
+ Major. Corps of Engineers, U. S. A. 



THK IMPROVEMENT OF THE OHIO RIVER 389 

miles, while the entire water-shed comprises an area of about 214 000 
sq. miles, which is larger than the area drained by the Mississippi River 
above the mouth of the Missouri, and larger than the drainage area of 
any other river in the United States, except the Missouri and the Lower 
Mississippi. The total fall at low water from Pittsburg to Cairo is 
•126 ft. In the lirst 26 miles of the river, from Pittsburg to Beaver, the 
average slope is about 15 in. per mile; from Beaver to Wheeling, a 
distance of 64 miles, 9 in. per mile; from Wheeling to Louisville, a 
distance of 509 miles, 5^ in. per mile; while from Louisville to Cairo, 
the slope is only 4 in. per mile. 

In its upper portion, the river is subject to frequent and rapid 
fluctuations of water surface. 

In connection with the generally accepted theory that the destruc- 
tion of the forests has increased the number and extent of floods, and, 
at the same time, decreased the low-water flow of the streams, the 
following statistics are cited: 

Records of the river stages in Pittsburg for the 24 years preceding 
1881, show the following: 

Average Number 
River Stage. of Days per Year 

i to 1 ft i day. 

1 to 2 ft 34 days. 

2 to 3 ft 37 days. 

3 to 4 ft 47 days. 

4 to 5 ft 43i days. 

5 to 6 ft 40 days. 

Average number of days when river was below 6 ft., 214. 
Average number of days when river was above 6 ft., 151. 

For the 24 years succeeding 1881 : 

Average number of days when river was below 6 ft., 201. 
Average number of days when river was above 6 ft., 164. 

In Colonel Ellet's book, "The Ohio and Mississippi Rivers," the 

following record of lowest water depths on the bar at Wlieeling is 
found : 

Sept. 27, 1838.. ft. lOHn. Sept. 19, 1845... 2 ft. 2 in. 

Sept. 6, 1841.. 1" " Oct. 13, 1846... 1" 9" 

Aug. 17, 1843.. 1 " 8 " Sept. 8, 1847... 2 " 3 " 

Sept. 23, 1844. .1 " IJ " Sept. 18, 1848... 1 " 11 " 



390 



THE IMPHOVKlVrENT OF THE OHIO RIVER 



The number of times that the Ohio River was above the danger 
stage at Pittsburg, Cincinnati, and Louisville, during the sixteen years 
preceding 1891 and the seventeen years following, is shown in Table 1 : 

TABLE 1. 



Place. 


Height of 

danger line above 

zero of gauge, 

in feet. 


First period, 16 

years, 
1875-1890, inclu- 
sive. 

Times. 


Second period, IT 
years, 
1891-1907, inclu- 
sive. 
Times. 


Pittsburg, Pa 


22 
50 
28 


18 
11 
11 


26 


Cincinnati, Ohio 


11 


Louisville, Ky 


9 







It is to be expected that the records at Pittsburg would show a 
greater number of river stages above 22 ft. in the years succeeding 
1890 than in the preceding years, because of the material contraction 
of the river channels at that place due to encroachments of railroads 
and manufacturing plants. However, the conditions at Louisville and 
Cincinnati have changed so little that fair conclusions can be drawn 
from the records at those places. 

While the highest gauge readings of which there is record in 
Pittsburg occurred March 15th, 1907, on which date a stage of 35,6 ft. 
was reached, it is thought that during the flood of 1832, which reached 
a height of 34.94 ft. at Pittsburg, the discharge of the river was 
probably equal to that of March 15th, 1907. A comparison of the old 
and present maps shows a marked contraction of the river channels 
near Pittsburg on account of slag dumped over the banks into the 
streams. 

It seems to follow from these data that the extremes as to flood height 
are not materially influenced by forests, but that the frequency of 
medium size and possibly destructive floods may be increased by de- 
forestation. These data further show that the deforestation of the Ohio 
water-shed has had practically no effect upon the extent of the low- 
water navigation seasons. 

A theory has been advanced that the low-water stages of rivers are 
less dependent upon springs than upon summer rains, first on one 
tributary and then on another. This has been especially noted on the 
Monongahela River, which stream becomes exceedingly low during the 
summer and fall months, unless there be local rains, notwithstanding 



Tin-: i.\rrit()VKMKNT of the ohio riveu 391 

a copious i)ri(>r winter and spriny rainy season. Sunnner rains, if the 
lands be denuded of forests, reach the river; whereas, if a forest exists, 
the rain is absorbed by the thirsty plant life, or evaporated. 

The maximium recorded discharge of the Ohio River just below 
Pittsburg, is 439 000 cu. ft. per sec. The discharge at Davis Island 
Dam, just below Pittsburg, has been observed in recent years to be as 
small as 1 GOO cu. ft. per sec. Captain Saunders gauged the Ohio River 
at Pittsburg in 1838, during the drought, and found the discharge to 
be 1 400 cu. ft. per sec. The quantity of water emptied into the Ohio 
by its tributaries during the dry season is very small, the low-water 
discharge of the Monongahela River being only 166 cu. ft. per sec. 
Although the low-water discharge of practically all the tributaries is 
too small, and the slopes of many of them are too steep to admit of 
an efficient improvement by regularization, the discharge is large 
enough to permit an extensive commerce to be carried on through a 
system of locks and dams, that on the Monongahela River being about 
12 000 000 tons per year. It is not thought that such an extensive 
use of this or other streams of similar slope would be possible unless 
the same were canalized, this assuming the existence of svifficient water 
to make tlie needed open-river ehannel depths. The present system of 
dams with movable tops will create a navigable depth of about 11 ft. 

Even during the dryest periods, an abundance of water exists in 
the Ohio for the maintenance of a 9-ft. depth in the pools of a slack- 
water system, as has been demonstrated at Davis Island Dam, built 
more than twenty years ago. 

Usually, the low-water period includes July, August, September, 
October, and November. During the low-water season, the depth of 
water on the bars near Pittsburg is from 12 to 18 in.; sometimes for 
nearly three months the bar depths are less than 2 ft. The lack of 
navigable depth in the lower river is caused largely by the excessive 
width of the channel. 

Table 2 shows the number of days, within the period named, that 
a 9-ft. stage or more existed at imi)nrtant ])oints along the river. 

The present project for improving the Ohio contemplates a naviga- 
ble depth of 9 ft., which project, without changing any structures, could 
easily provide for 11-ft. navigation by dredging in the upper end of 
the pools. It will be observed from Table 2 that the number of days 
per year during which there is a natural stage of 9 ft. or more, in- 



392 



THE liMPliOVEiVIENT OF THE OHIO JIIVER 



creases as the river is descended. From Pittsburg to Beaver, the mean 
number of days that a 9-ft. navigation is afforded is eighty-one. The 
Pittsburg District, including a radius of 50 miles around the city 
proper, is the most prolific freight-producing center in the world. The 
intermittent navigable stages in the river, aggregating only 81 days 
per year, has debarred this district from an extended use of the Ohio, 
(.'xcept for the cheaper grades of freight, such as coal, and for that only 
where a long haul is necessary. 

TABLE 2. 



Place. 


Miles. 


Mean number of days at or above 
9 ft. during the 10 years, 1895- 
1905. 


Pittsburg, Pa 


0.0 
28.5 
90.0 
183.5 
263.4 
353.0 
466.5 
553.5 
599.0 
783.0 
919.5 
967.0 


81.0 


Dam No. 6 


117 


Wheeling, W. Va 

Parkersburg, W . Va 


119.2 
140 5 


Point Pleasant, W. Va 


150.3 


Portsmouth, Ohio* 

Cincinnati, Ohio 


216.4 
248 3 


Madison, Ind.* 


233 


Louisville, Ky 


97.5 Head of Falls. 


Evansville, Ind 


198 5 


Paducah, Ky.* 


203.6 


Cairo, 111 


302.4 







* Five years' records only. 

During the periods that intervene between freshets, a great quantity 
of freight is kept in the harbor at Pittsburg awaiting a rise. The 
steamboats are tied up, much capital is idle, and great expense is in- 
curred in watching the boats and barges, pumping them out, etc. 

"In June, 1895, there were collected in the Pittsburg harbor, 
1 200 000 tons of coal, loaded upon about 2 500 vessels awaiting water 
to move them down the Ohio River, the largest tonnage ever assembled 
in any harbor of the world at any one time. The rise did not come 
until November 27th. The cost of freight and vessels engaged in this 
service was estimated at $6 310 000. It cost $2 000 per day to keep 
the tonnage afloat, and $1 000 per day interest on the investment. 
Total, $3 000 per day. This tonnage was kept waiting in the Pittsburg 
harbor for water in the Ohio River an average time of five months, or 
150 days, at a loss of $450 000, which is 5% of $9 000 000. This shows 
what one item of commerce lost in five months, because Western Penn- 
sylvania did not have the economy of transportation that results from 
continuous water movement."* 



* Adtlress by Mr. John E. Shaw, before the Merchants' and Manufacturers' Association, 
at Pittsburg. 




fLATE XX. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 



STEAMER "SPRAGUE" AND TOW 

LOUISVILLE TO NEW ORLEANS. 
PRESENT OPEN-RIVER NAVIGATION 



STEAMER SPRAGUE" 
Length, 315 feet 




Tow: 50 Coal Boats and 2 Fuel Boats. 

Containing 50000 Tons of Coal. 

Area covered by the SO Coal Boats 5.9 Acres. 



Note: 

System of tying Boats together and to Steamer 
is indicated by broken lines. 
Draft of coal boats 8 to SJi feet. 



-Length of Steamer "Sprague" and Tow 1132 feet 



THE l.Ml'UOVK.MKN r OF THE OHIO RIVER 393 

Boats and Commerce. 

Boats suitable for navigating the Ohio River do not anchor, but are 
tied to the shore, and a sloping shore is needed so that they may be 
moved toward the bank as the vpater rises, thus keeping out of the 
stronger currents where drift would lodge against them and break them 
loose. 

The short-lived freshets of the Upper Ohio have developed an 
unique system of towing on tlie Ohio and Mississippi Rivers. At 
times of freshets, freight is carried in great fleets made up of boats 
and barges. The boats are about 26 ft. wide by about 175 ft. long, 
and each carries approximately 1 000 tons. The barges are 26 ft. 
wide by about l;>0 ft. long, an<l each carries about 500 tons. As they 
leave the upper river, the fleets vary in size, containing from 10 to 20 
boats or barges. These fleets are -increased in size as they proceed down 
the river, and sometimes comprise from 50 to 60 boats of 1 000 tons 
each, the fleet covering 5 or 6 acres of water. 

Plato XX shows the Steamer Spragne and a tow of coal barges. 

When loaded, coal barges draw 6 ft., and coal boats from 8 to 9 ft., 
the former requiring a navigable depth of 8 ft., and the latter from 
10 to 12 ft. The excess depth of from 2 to 3 ft. is needed, because 
the large fleets generally require a channel width of at least 300 ft. of 
the full depth of the deepest laden craft in the fleet, and to obtain this 
it is generally necessary to have a river stage at least 2 ft. higher than 
that necessary for a single boat. A slack-water depth of 9 ft. is at 
least e<]ual to an 11-ft. open-river depth. With slack water, the pools 
are more than 9 ft. deep at all places except immediately below the 
locks. There being no currents to interfere, a little dnnl^iiia- allows 
fleets of full depth and width to pass into the deeper water and wider 
channels further down in the pools. The absence of currents greatly 
assists navigation in keeping in narrow channels. 

The steamboats are of the stern-wheeled type. Experience has 
shown that these boats are exceedingly efficient for the work required. 
They are equipped with long balanced rudders against which the water 
from the wheel impinges with high velocity when the boats back, thus 
driving the stern in the desired direction whether the boat has head- 
way or not. With this device, boats of 2 000 h. p. can successfully 
maneuver, down stream in swift water, fleets of from 40 000 to 60 000 
tons burden. 



394 THK IMPROVEMENT OF THE OHIO RIVER 

In a report submitted to the Chief of Engineers, in 1870, tiie late 
W. Milnor Roberts, Past-Px-esident, Am. Soc. C. E., stated that: 

"A system which will provide uninterrupted navigation will un- 
doubtedly revolutionize the system of coal carriage. Less powerful 
steamers can then be used steadily all the year round, making regular 
trips, each boat taking fewer barges, but doing the same, or more busi- 
ness, with less capital invested at less risk, and without the expense 
due to idle boats and the expense of watching the great fleets while 
awaiting a freshet." 

A system of locks and dams was referred to by Mr. Roberts in the 
paragraph quoted. 

The records of freight cost under the present towing system, to- 
gether with the estimated cost with a slack-watered Ohio, indicate 
that the export freight of the Ohio Valley, offered in barge-load lots, 
should be delivered at New Orleans at about $1 per ton.* Boats could 
then advantageously transport freight up and down stream. In an 
unimproved river, or in one improved by regularization, currents are 
often a serious handicap to up-stream navigation. 

By adding to the foregoing cost a reasonable water rate from New 
Orleans to San Francisco, including a toll charge of 75 cents per 
ton through the Panama Canal, it appears that, with a slack-watered 
Ohio and a completed Panama Canal, freight should be transported 
from the Ohio-Mississippi water-shed to the Pacific Coast at from 
$5 to $8.50 per ton. With the present class rates by rail from Chicago 
or New York to San Francisco at from $19 to $32 per ton, and with 
commodity rates at about 4.3 mills per ton-mile, one would expect a 
large commerce, via the all-water route, between the people of the 
Ohio-Mississippi water-shed and those living between the Rockies and 
the Pacific Ocean. The Panama Canal will, of course, offer similar 
opportunities to the Atlantic Coast, both as to trade with the Pacific 
Coast and with China and Japan. In this connection the following 
is inserted : 

"Recently 2 .500 tons of stool mils were shipped from Pittsburg for 
the constniclidii of the Prince Kupcrt niid of the (irniid 'rniiik I'ncitic 
Systciii, l>,v wliiil, on ils fiu'c ni'ponrs to bo an oxceodiiigly nmiKhilionl 
route. The mils were HrsI sent to Now York, where tlu-y were lo;i(le<l 
on vessels going to the Oi-ieiil by the Suez Ciiiial Route. On reaching 
Kobe they will be transshipped and sent across the Pacific to thoir 

♦Executive Doc. 492, 60th Cong., 1st Sess., pp. 88 to 35. 



PLATE XXI. 

TRANS. AM. SOC. CIV. ENGRS. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 




g «J...yc 




^Montreal./ i \\ ( ,. 
.i. / <«. trToi'tMuonth 

^-^y^oiC — Il1i"'_l 

- '^'— "^N^w YORK JbRALTAR 3202 M 



^ 






\ 



Lh.S. BOAR.D-OF ENGINEERS \ 
'~^ on lmpj:ovement of the Ohio River, \ 
(Constituted'by S.0.17, Office, Chief of Engineers, 
, ■''''' May 12,1905) \ 

fe^/'^'ShcwIng territory commercially benefited 
by reliable navigation ^ 

OHIO AND MISSISSIPPI SYSTEMS 

100 '6 60 25 ICO 2C0 

SCALE OF MILES 



Till:; IMIMJOVK.MENT OF THE OHIO KIVEH 395 

final destination. ITow much will be saved by this method of shipment 
is not stated." 

It is thought that the tolls charged on the Panama Canal should 
be low — not much more than enough to cover the operating and mainte- 
nance expenses. The construction cost should be charged largely to 
National protection, at least until a large traffic is developed. 

'i'lic iiiai>, Plate XXI, accompanying a report of an examination 
of the Ohio River made by a Board of Engineers, United States Army, 
and transmitted to Congress in January, 1908, shows the territory 
that it is thought will be commercially benefited by reliable navigation 
in the Ohio-Mississippi system. 

The Report states :* 

''The boundary line of this territory was found by drawing a line 
through tow^ns to which the freight rate from New Orleans by river 
and rail at present rates is equal to the present all-rail rates from 
seaport towns to the same points. Barge freight rate from New 
Orleans to distributing points such as Vicksburg, Cairo, Evansville, 
Louisville, Cincinnati, Wheeling, and Pittsburg was assumed at 1.14 
mills per ton-mile, which was the rate charged during the year 1905-6. 
The rail rate used was 4.832 mills per ton-mile from Gulf cities north- 
ward and is the present mean ton-mile rate on sugar in carload lots 
from New Orleans to Chicago, Cincinnati, Louisville, and Pittsburg. 
The rate used from Atlantic Coast cities inland is that of the Penn- 
sylvania Railroad from New York to Pittsburg, or 7.4 mills per ton- 
mile, which is the mean of the ton-mile rate on sugar and rice. A rate 
of 2 mills per ton-mile was used on the Hudson River and through the 
Erie Canal, and a rate of li mills per ton-mile on the Great Lakes for 
such commodities as sugar, rice, etc. 

"Within the territory shown on the map now lives a population of 
about 20 000 000 people, 8 504 000 of whom live in that portion of ter- 
ritory lying in the Ohio Valley proper." 

jMetiiods of Improvement. 

The original method of improvement of the Ohio River included 
the "removal of snags and rocks; the closing of duplicate channels by 
low dams; the guiding of the water into a single low-water channel 
by dikes where the river had an excessive width; and the removal, by 
dredging, of hard bars and projecting points." 

It was only hoped by this method to prolong the navigation season 
for packet business, enough water not being available for producing 



♦Executive Doc. 492, 60th Conp.. 1st Sess.. p. 22. 



396 THE IMPROVEMENT OF THE OHIO RIVER 

a depth and width sufficient for towboats and fleets. Economy of 
transportation on the Ohio being within limits proportional to avail- 
able depth and width, experience indicated that fleets from Pittsburg 
to New Orleans should draw at least 9 ft., and that channel widths of at 
least 300 ft. were necessary at the lowest navigable stage. This naviga- 
ble condition could only be brought about permanently by materially in- 
creasing the low-water flow of the river, or by locks and dams. 

The following methods of increasing the low-water discharge have 
been proposed: 

(1). — To obtain the additional water needed during the rainy 
season by cutting a ditch from Lake Erie to the Ohio River. How- 
ever, as the river above Marietta, Ohio, was at a higher elevation 
than Lake Erie, this method was impossible. 

(2). — That Lake Chautauqua should be used as a regulating reser- 
voir for the Ohio River. The drainage area of the lake was too small, 
however, for serious consideration. It was then proposed to use Lake 
Chautauqua as a reservoir, and to keep it filled by pumping from 
Lake Erie. 

(3). — Colonel Charles Ellet, a civil engineer engaged in work 
at Wheeling, West Va., made some observations of the discharge of 
the river there, from which he reached the conclusion that the average 
annual discharge of the river was sufficient to maintain a continuous 
6-ft. navigation, if the water was stored in reservoirs at times of large 
discharge and permitted to flow at times of small natural discharge. 
The plan proposed involved the construction of large dams to form 
reservoirs upon the head-waters. This scheme was reported upon ad- 
versely by W. Milnor Roberts, United States Civil Engineer, who was 
in charge of the Ohio River improvements from 1866 to 1870. Mr. 
Roberts was peculiarly well qualified to investigate the subject, both 
because of his efficiency as a civil engineer, and also because of his inti- 
mate knowledge of the region drained by the Allegheny and Mononga- 
hela Rivers, he having been connected at different times with the 
Pennsylvania State Canal between Johnstown and Pittsburg; the 
canal between Erie and Beaver; the Monongahela slack-water system; 
and surveys for various railroad lines in the regions mentioned. 

This plan has recently been revived. It should be remembered, in 
considering it, that an open-channel depth of at least 11 ft. is now 
necessary, the trend being always for deeper water. A 14-ft. depth 



THE l.Mrii()\ HMENT OF TlIK OHIO KIVKU 397 

is being advocated between Chicago and St. Louis. The records seem 
to indicnk' tli;it the supply of water is not sufficient to provide such a 
depth during the entire year in the Upper Ohio Kiver, while there 
may be sufficient for the Lower Ohio. 

The proposition to increase the low-water flow by means of water 
impounded in reservoirs, had, of course, the double object of thus 
improving navigation and, at the same time, providing for flood pre- 
vention. The extreme irregularity with which floods occur and the 
navigation necessity of always having full reservoirs at the beginning 
of the dry season, would result many times in the reservoirs being 
full when needed for flood prevention. Thus they would not be able 
to restrain the flood in such manner as to prevent entirely the usual 
damage. 

If flood prevention alone be the function of the reservoirs in the 
Upper Ohio, the writer believes that they could be made to serve such 
purpose. During the winter season, when the movable dams were down, 
the impounded water might be used to prevent the formation of ice 
in large quantities, by increasing the flow so as gradually to drive out 
such ice as forms, thus acting as an auxiliary to the present movable 
dam system. The water of the Chagres River, on the Isthmus of 
Panama, is to be impounded, in order to control its flow, in a reser- 
voir which will be 165 sq. miles in area. The flood discharge of this 
river is about one-third that of the Ohio at Pittsburg. 

Reservoirs as means of river improvement have always been attrac- 
tive. Their cost, however, in connection with the danger of their 
being filled with detritus from the surrounding hills, has prevented 
their adoption generally as a system of river improvement, both in the 
United States and in Europe, and that on streams where the slope was 
such that commerce could obtain the full benefit, if the necessary depth 
existed. 

A company was organized in 1855 for the purpose of experimenting 
on the Ohio River with a system of improvement which had been 
devised by General Herman Haupt. The Company proposed to prose- 
cute the work at its own expense and on condition that no payment 
should be made by the Government unless the plan proved to be a suc- 
cess. In brief, the plan of improvement involved the use of low dams 
and open chutes, about 300 ft. wide, inclosed on one side by the 
natural bank of the river and on the other by a longitudinal bank or 



398 rUE IMPROVEMENT OF THE OITTO KIVEK 

mound. This plan was reported upon adversely in 1870 by Mr. 
Roberts, and again in 1880 by a Board of Engineers. 

Mr. Alonzo Livermore proposed that low dams be used, but that, 
instead of locks, long chutes should be provided. These chutes would 
contain division walls and enlargements located so as to decrease the 
velocity and the discharge of water through the chutes. 

Mr. Roberts investigated all the above-mentioned plans and then 
decided to recommend that the river should be canalized, locks and 
fixed dams being used, each dam being provided with a navigable 
chute for use during freshets of moderate height. 

A few years later, the late W. E. Merrill, M. Am. Soc. C. E., Major, 
Corps of Engineers, United States Army, investigated the subject of 
a radical improvement of the river and agreed with Mr. Roberts that a 
lock and dam system was the proper solution of the problem. He 
stated that the advantages of a canalization scheme over any other pro- 
posed method of improvement were as follows : 

(1). — It has been long tried and is now in use on the Mononga- 
hela River, where it meets the demands of the same commerce that 
navigates the Ohio. 

(2). — There are no great hazards connected with the system, since 
the dams are low and the destruction of one would not necessarily 
injure the one next below. 

(3). — It is known iiositively that locks and dams can be built that 
will answer the purpose fully, and their cost can be determined 
beforehand with very fair accuracy. 

(4). — There would be no damages from overflow, or destruction of 
property of any kind. 

(5). — No special care is needed in the use of the slack-water system. 
The pools are themselves reservoirs containing the minimum quantity 
of water needed, and at the exact place where it is to be used. 

(6). — The cost of the system would probably be less than that of 
the other (reservoir) system. 

(7). — The pools would make excellent harbors for all river craft, an 
improvement that is greatly noodod nt tlic large cities, and especially 
at Pittsburg. 

There was one great objection, however, to the construction of any 
system of dams such as had ever previously been built in the United 
States. At times of freshets, all the coal shipments were made in fleets 



THE IMPROVEMENT OF THE OHIO lUVKU 399 

which were too large to go through a lock of any reasonable dimen- 
sions, without being divided into two or more lockages. When the 
fleets are made up, the barges are fastened together very securely by 
cables in every direction, and the delays which would be caused by 
making iuui uiiinakiiig a Hfct at each lock would be very great. 
(Plate XX.) The coal shippers were vigorously opposed to the con- 
struction of any system of dams which would not permit the passage 
of fleets down stream, at times of freshets, without being broken up 
and reformed at each dam. It was believed that the influence of these 
interests was such that they would be able to prevent the appropriation 
by Congress of funds for any system of dams which would change the 
established system of coal carriage during high stages of water. 

A Board, consisting of Majors Weitzel and Merrill, then made a care- 
ful study of the whole subject. A system of fixed dams was studied first, 
each dam to have an opening, 4 ft. deep and 200 ft. wide, cut in the toj), 
with an inclined plane below the dam, and built so that, at high stages, 
coal fleets could be passed through the opening and down the inclined 
plane into the lower pool. The opening in the dam was to be closed by 
some form of movable gate which could be operated easily and quickly. 

The Board studied the French and other systems of movable dams, 
but considered that there might be objections to the use of Chanoine 
wickets on the Upper Ohio, which did not exist in the case of the 
rivers of France. It was believed that the large quantity of drift and 
ice found in the Upper Ohio might be very injurious to the parts of a 
movable dam of the French type. Major Merrill's plan was finally 
adopted in part. This plan provided for 13 locks and dams between 
Pittsburg and Wheeling, built for a minimum depth of 6 ft., the 
dams to be of the movable type, with navigable passes 400 ft. wide, 
to be closed by Chanoine wickets, and high and low weirs, to be 
closed with either Desfontaines' wuckets or Brunot's gates; the locks 
to be 78 ft. wide by about 630 ft. long, with lifts of from 4 to 7 ft. 
The first dam was located at Davis Island, five miles below Pittsburg. 
The plans for the first lock and dam wore subsequently modified so 
that the lock as built was 110 ft. wide and 600 ft. long, and the weirs 
were closed by Chanoine wickets of shorter lengths than those of the 
pass. All wickets were originally planned to be maneuvered from 
service bridges. Davis Island Dam was commenced in 1878, and 
opened to navigation in 1885, and its cost was about $910 000. 



400 THE IMPROVEMEMT OK TITE OHIO KIVEK 

The operation of the dam at Davis Island having been successful, 
in 1888, Major Merrill proposed the immediate construction of two 
more dams of the series. The Board of Engineers went still further 
and recommended the extension of the slack-water system as far as the 
mouth of the Beaver Eiver, by the construction of four of the dams 
previously mentioned in Major Merrill's project of 1874. The first 
appropriation for building Dam No. 6 was made in 1890 and the last 
in 1902, and the lock and dam were completed and put in commission 
in the summer of 1904. 

The first appropriation for the construction of Locks and Dams 
Nos. 2, 3, 4, and 5, was made by Congress in June, 1896, and the 
last in 1905. These latter dams were completed in 1906 and 1907. 
The time taken by Congress to appropriate the money for these dams 
is specially mentioned, because engineers are often blamed for the 
slow building of structures for which legislative bodies have not pro- 
vided the funds, for which latter action engineers are in no way 
responsible. 

In the construction of such a system of improvement as described 
in this paper, the building of each lock and dam complete should be 
provided for by one contract. This would prevent the delays due 
to entering into many contracts at different periods of the construc- 
tion, and those due to collecting and erecting different sets of con- 
tractor's plant, etc. 

Each lock and dam complete should be built in from 3 to 5 years, 
and there is no limit, of an engineering nature, to the number of locks 
and dams that can be contracted for at the same time. Of course, this 
means that all the funds necessary to complete a lock and dam must 
be authorized prior to the commencement of the work. 

The River and Tlarhor Bill, passed in the sj)ring of 1905, authorized 
changes to be made in Locks and Dams Nos. 1 to 6, so as to provide 
for a navigable depth of 9 ft., instead of 6 ft., as originally planned. 

The harbor of Pittsburg had been dredged so as to permit naviga- 
tion by boats drawing between 9 and 10 ft., and, therefore, it was only 
necessary to change the plans of the dams which had not been com- 
pleted, viz., Nos, 2, 3, 4, and 5. The changes made comprised the 
raising of these four dams by means of longer wickets and higher 
bear-trap gates, and the lowering of the lower lock sill at Dam No. 5. 
No changes had to be made in the completed dams, Nos. 1 and 6, but, 



THE Il\rPl{OVKMHNT OF 'I'lll-: OHIO VAWM 101 

in order to permit of the extension of the harbor of Pittsburg to include 
a good natural deep-water pool a short distance below Dam No. 6, the 
lower lock sill at that dam was lowered. This will allow coal boats 
to be taken from Pittsburg during the navigable season, at any stage 
of the water, down as far as this good natural harbor, there to await 
a coal-boat freshet. 

The Act of Congress, approved March 3d, 1905, made provisions 
for a survey of the Ohio River from Pittsburg to Cairo, with a view 
to the radical improvement of the river throughout its entire length. 
Tlic report of the Board of Otiicers appointed to make the survey lias 
been submitted to Congress. This report recommends the canalization of 
the river from Pittsburg to Cairo, so as to provide a navigable depth of 
9 ft.; the locks and dams are to be built of practically the same type as 
thos(^ already constructed in the upper portion of the river; the Louis- 
ville and Portland Canal is to be widened, and a duplicate lock is to be 
placed therein. The estimated, cost of the entire project was $63 731 488, 
in addition to the appropriation already made, viz., $9 281 370. 

The locks recommended are to be 110 ft. wide and 600 ft. long in the 
clear. The dam is to consist of a navigable pass from 600 to 700 ft. 
wide, an automatic part, probably of the bear-trap type, and a series 
of movable weirs of the Chanoine, or other well-tried type; the propor- 
tion of weir and automatic part is to be determined by the conditions 
existing at each dam. 

Thus far the order of building has been first to construct locks 
and dams below the important cities and tributaries, so as to provide 
good harbors as soon as possible with the funds available, and thereby 
enable river commerce to connect with existing railroads. Thus, Davis 
Island Dam was built below Pittsburg; and, at present. Dam No. 13, 
below Wheeling, No. 18, below Marietta, and No. 37, below Cincinnati, 
are under construction. 

The following rules were promulgated by W. II. Bixby, M. Am. 
Soc. C. E., Major, Corps of Engineers, U. S. A., for the location of 
locks and dams in the upper river: 

"1st. To allow room in the river for a navigable pass at least 600 
feet wide, with its centre nearly in line with the existing channel at 
tow-boat stages of water. 

"2nd. To allow room in the river for weirs of at least 240 feet total 
length, so placed as to be as much as possible protected from injury by 
floating drift. 



402 Till': i^iriiovEMENT of the oiiio river 

"3rd. To allow room in the river for passage of coal tows around 
each half of the dam during construction of the other half. 

"4th. To allow of the sill or bottom of the navigable pass being 
placed at least slightly above the present bed of the river, and yet at 
least as far below the water surface as are the channel bars, which at 
coal-boat stages determine the available depth of the river in the pools 
above and below. 

"5th. To avoid positions where the coal-tows must be flanked or 
placed obliquely to the channel line by reason of river bends, or oblique 
currents. 

"6th. To allow of the view of dams by approaching boats for dis- 
tances of at least half a mile both up and down stream. 

"7th. To avoid positions near shifting sand or gravel bars, and 
mouths of sand-bearing rivers, creeks, and runs, unless such bars and 
streams be below the dams on the weir side. 

"8th. To allow of positions below the mouths of large tribiitaries 
like the Little Kanawha, Great Kanawha, Guyandotte, Big Sandy, 
Scioto, Licking, and Big Miami, near enough to extend navigation 
properly to the nearest dam in such tributary, but yet far enough below 
the mouth of such tributary to avoid trouble from deposits and flood 
overflows, and to allow the proper handling of boats entering or leaving 
such tributaries. 

"9th. To avoid cutting up or otherwise injuring the existing 
harbors of large cities, especially Parkcrsburg, Pomeroy, Point Pleasant, 
Gallipolis, Huntington, Catlettsburg, Ashland, Ironton, Portsmouth, 
Cincinnati, Covington, and Newport. 

"The first, second and third conditions prevent locations in narrow 
parts of the river, and especially at islands where the bank channel is 
bare at low water, and of insufficient depth at coal-boat stages. The 
third condition prevents location within a half mile of the ends of 
islands, or high mid-river bars. The fifth and sixth conditions prevent 
locations at sharp bends. 

"While it is not to be expected that all of these conditions can be 
fulfilled at each location, still past experience on this river, combined 
with the information so far secured, indicates that these conditions 
can, in general, be complied with, and that departures from the same 
will be the exception rather than the rule." 

It is important, in planning any system of slack-water navigation, 
to foresee and provide for any increased depth which may be demanded 
in the near future. It is better to design the locks so as to accom- 
modate a greater draft than that indicated at the time of construction, 
in order that increased depth may be easily obtained in the future by 
dredging in the pools. The extra cost of a lock so built is limited to 
the excess cost of the lower gate over the decreased cost of the lower 



THE IMPROVEMENT OF THE OHIO RIVER 403 

sill foundations. For this reason, any system of locks and dams should 
be designed so as to provide the draft needed, without the use of aux- 
iliary dredging, leaving the deepening of the pools to be done later, 
to provide for a possible demand for increased depth. 

It will be remembered that the dams originally proposed for the 
Ohio were fixed dams provided with a navigable chute. The require- 
ments- of the coal-fleet navigation made fixed dams which were not 
provided with some kind of navigable pass out of the question. 

A movable dam to be used on the Ohio River should conform to 
the following: 

It should be easily operated and free from complicated mechanism. 

It should be capable of being lowered rapidly, without chance of 
failure. 

Its parts should be so strong and simple as to prevent damage by 
ice and drift. This consideration is of very great importance on the 
Ohio, where the severity of the winters must be kept in mind in select- 
ing a type of movable dam. 

It should be capn])]e of being easily operated, so as to i)!iss small 
freshets without the necessity of lowering the entire dam and without 
lowering the pool below normal level. 

It should afford a navigable channel at least 600 ft. wide, with the 
dam down, and the passage should then have a depth of water equal to 
that on the bars above and below the dam. It is thought that the 
combination of Chanoine pass and weirs, with bear-trap automatic 
portion, fills these conditions better than any other. 

The lift of the dams planned, up to the present, has been limited to 
about 8 ft. as a maximum. This was originally based on the experience 
in France. With these small lifts and a navigable stage of only 6 ft., 
the wickets for the navigable pass were of a length not greater than 
14 ft. The increase in navigable depth on the Ohio, from 6 to 9 ft., 
caused the use of wickets as long as 18 ft. These have been maneuvered 
satisfactorily, and it is possible that Chanoine dams of a 10-ft. lift 
could be satisfactorily operated. 

A wicket 16 ft. long was adopted in the project last recommended 
to Congress. If the construction engineers can safely lengthen these 
wickets, a smaller number of dams will be needed. 

There are no lift walls in the locks, both gate-sills being at ap- 
proximately the same elevation as the sill of the pass. This is advan- 



404 THE IMPKOVEJMKA'T OF THE OHIO UIVER 

tageous, because it permits boats to go through the lock at times when 
the dam is down and there is not much water in the river. When an 
upward-bound packet meets a tow going down stream through the pass, 
it is often desirable for the packet to pass up stream through the lock. 

There are no guard walls prolonging the river wall of locks, because 
it has not been believed that they are necessary. 

Where the foundations of the river walls of the lock are such that 
no serious scour may be anticipated on the outside, due to the opera- 
tion of the dam, it is customary to let the dam start from about the 
center of the lock wall, thus affording ample room for emptying and 
filling the valves, which are placed in the outside wall itself. 

Whenever there is danger of scour at the place designated above, it 
is customary to let the dams abut against the wall as near the lower 
end of the locks as possible, and, at the same time, to provide the 
necessary filling and emptying arrangements. Eddy currents below 
movable dams should not be very serious, since the dams are down at 
all higher stages of the river. 

Before a sudden rise, it may be essential to commence lowering the 
navigable pass at the end next to the lock, although this should not be 
done unless absolutely necessary. If it is done, the scouring effect of 
the current alongside the lower end of the river wall is very great. It 
should not be concluded that, because movable dams are down during 
the higher stages, scour is not probable. 

In case of accident to the lock gates, the rush of water through the 
lock would probably undermine the walls, imless they were properly 
protected against scour. Moreover, during construction, there are times 
when the cross-section of the river at the site of the dam is very much 
reduced by coffer-dams, etc., and the currents through the openings 
between obstructions are such as to produce very dangerous scour. The 
wall foundations should be protected against any of these contingencies. 

At Lock No. 4, during construction, the entire width of the river 
was at one time closed, except about 430 ft. which was divided into 
two channels. One of these channels, next to the lock wall, was only 
220 ft. wide. The scour along the lower end of the river wall, caused 
by the severe currents through this narrow passageway, threatened to 
undermine the lower end of the wall, and the scour was only stopped 
by the prompt placing of a large quantity of rip-rap, thrown by hand 
from the top of the wall. 



THE IMl'liOVKMKXT OF THE OHIO IlIVEK 405 

The principle which should govern the height of the walls above 
the pools is that they should be of such a height as will permit a boat 
to lie alongside them with its guard below the top of the wall. This 
wuuld indicate that 5 ft. above the water is the proper height to ac- 
couniiodate packet-boats, the guards of which arc generally from 4 to 4^ 
ft. above the water surface, but, of course, this depends on the load 
carried. Towboats have much lower guards. As the lock may be used until 
the upper pool rises to a height of about 1 ft. above the crest of the dam, 
it would be better to build the upper guide wall and the land wall of 
the lock to an elevation of about 6 ft. above the level of the normal 
upper pool. The river wall need not be built so high, because it is 
unusual for packets to lie alongside that wall while making a lockage, 
and, if necessary, their fenders can be used at any time. In fact, the 
dams in the stretch of river near Wheeling have river walls of less 
elevation that the land walls; but, as originally built, the walls of the 
locks in the upper part of the river were uniformly 5 ft. above normal 
pool level. The increase in pool elevations, caused by the change from 
a 6- to a 9-ft. navigable depth, decreased the height of the walls 
above the pools and necessitated raising some of the walls. In most 
cases, the land wall and lower guide wall were raised, the river walls 
not being raised. The lower guide walls were originally built to an 
elevation of 5 ft. above normal lower pools, which provided the same 
clearance as the other walls, when the pools were at normal level. 
However, this did not take into consideration that when there is con- 
siderable flow over the dam, the lower pool rises almost twice as fast 
lis the ujipcr pool. \\ would seem then that, for a symmetrical design, 
the lower guide walls should be built to an elevation of about 7 ft. 
above normal lower pool, if the land wall is to be 6 ft. above normal 
upper pool. A guide wall should be at least as long as a tow that 
can pass through the lock in one lockage. In the latest dams on this 
river, the guide walls are from 600 to 700 ft. long. 

Lock-Gates. 

The Ohio Eiver locks are differentiated from most other locks in 
the United States principally by the type of lock-gate used. The 
gates are of the rolling type designed by the late Major W. E. Merrill, 
When out of use, a gate of this type is housed in a recess in the bank; 
when needed, it is run across the lock, thus closing it, the screen serv- 



406 THE IMPROVEMENT OF THE OHIO RIVER 

ing as a gate. As originally planued the locks were to be 78 ft. wide. 
It was not believed to be practicable to make tiiem wider, because ot 
the small height and the great length of the miter-gates. However, 
when it was deemed necessary, later, to increase the width to 110 ft., 
a form of gate suited to such a wide span and small height was de- 
signed. These rolling gates run on tracks in much the same way as an 
ordinary railway car. The principal parts of a gate of this type are : 

First. — The top truss which serves to carry a portion of the water 
pressure to the lock walls. 

tSecond. — The vertical water-screen which serves to close the lock 
and transfers one-third of the pressure to the top truss and two-thirds 
to the track. 

Third. — The trucks, wheels, etc. 

The original gates at the Davis Island Lock were made of pine 
timber. The top truss was of the Howe type, supported on posts so 
as to lie above the normal level of the upper pool. The posts extended 
below the level of the track, and their lower ends served as flanges for 
the wheels, which were of the plain-tread type, traveling on a Hat-rail 
track, the gauge of which was 11 ft. 6 in. The spacing of the rows of 
posts was such as to provide for a 2-ft. lateral movement of the gate; 
for the gate to move laterally, however, it was necessary for it to slide 
on the wheels parallel to the axis of the lock. This sliding created 
great strains on the lower part of the gate, and resulted in the breaking 
of many wheels and axles. The gate was moved back and forth by two 
chains, each chain being fastened at one end to an end of the gate and 
wound around a chain drum at the entrance to the recess. The chains 
were rigidly attached to the gate, and this mode of fastening was found 
to give much trouble. If the gate struck a submerged snag or rock on 
the track, it stopped moving instantly, and something broke before 
the engine could be stopped. Wheel axles were broken several times, 
and, on one occasion, when the gate could not be moved out of the 
recess, it was found that the outer axle was broken, the lower wheel 
of the second axle off, and the axle out of its bearings. 

After the original wooden gates had served for 12 years they were 
replaced by steel gates. In these new gates the top truss is of the Pratt 
type, the wheels are flanged, and a lateral movement of the gate is pro- 
vided for by hanging the framework upon each axle by eye-bar hangers, 
which allow the frame to swing laterally like a pendulum. This ar- 



PLATE XXII. 

TRANS. AM. SOC. CIV. ENGR8. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 




I'll.. 1. LdCK N(i. II. Ohio River. 




Kifi. 2.— (iATE Recess. Lock N'o. (J. Ohio Rivek. 




Fig. 3. — (Jate for Lock No. ti. « Hiio Kivkh. 



THE liMl'KOVK.MENT OF THE OHIO laVEK 407 

raiigemeiit prevents the friction of the bottom bearing from interfering 
with the opening of the gate. When the pressure is on tlie gate, it 
swings down stream about 2 in. until the bottom-bearing piece rests 
against the up-stream side of the down-stream rail. When the pressure 
is released, the gate swings back so that the bearing strip is about 2 in. 
away from the rail. In the new gate, rectangular valves are used 
instead of the circular butterily valves of the original gatos. 

The latest gates are those at Locks Nos. 2, 3, 4, 5, and 6. They 
are of steel, and are somewhat similar to the steel gates at the Davis 
Island Lock, The top truss is also of the Pratt type. The water-screen 
consists of horizontal white oak planks bolted to vertical, 15-in., 42-lb. 
I-beams, which rest, at the top and at the bottom, against horizontal 
members riveted at each end to the down-stream posts of the gate. 
Below the intermediate horizontal member and the bottom water-coal 
strip, are the gate valves. There are two of these valves in each 
panel, making eighteen in each gate. The valves are rectangular, of 
the horizontal-axis, butterfly type, approximately 3 by 4 ft. in size. They 
are made of structural steel closing against oak cushions, and are 
operated by hand from the top of the gate by racks and pinions. The 
valves were not constructed to be operated by power, because the locks 
were provided with filling and emptying valves in the river wall, and 
the gate-valves are not to be used except in case of accident to the wall- 
valves, and possibly to assist in flushing deposit out of the lock cham- 
ber and the tail-bay. (Plate XXII.) 

The pendulum arrangement for the lateral swing of the gate is 
much the same as in the later Davis Island gate already described. The 
connection between the superstructure of the gate and the trucks is 
through standard car springs which serve to prevent shocks to the gate 
when the wheels run over obstructions on the track. The wheels are 
standard car wheels, slightly modified in the tread so as to fit the level 
gate-tracks. The water-seal at the bottom is formed by a wooden bear- 
ing strip. The water-seal at each end of the gate is formed by a pivoted 
pipe arranged so as to be swung by the pressure of the water until it 
closes the opening between gate and wall. The operating chains are 
IJ-in. crane chains fastened at each end of the gate to a tug-lever, 
which is arranged with spring control so as to diminish the shock due 
to starting and stopping the gate. In order to prevent breakage of the 



408 TIIK I.MPKOVKMEiXT OF Till-: OHIO RIVER 

chain, the connections to the tug-levers are made by shackles and pins 
of such a size that the pins will break before any link in the chain can 
possibly fail. The gate is moved back into the recess and out across 
the lock by the rotation of the chain drum at the entrance to the recess. 
This drum and its shaft are rotated through a system of gearing by a 
10 by 12-in. engine of the horizontal, double-cylinder, reversible type, 
operated by compressed air, at Locks Nos. 2 to 6. At Davis Island a 
12 by 16-in. steam-driven engine is used. 

In the design of these gates, no estimate vpas made of the stresses 
v^fhich would result in case a gate should be struck by a boat entering 
the lock. All truss and screen members were designed to support the 
static water load, under the assumption that the lower pool had been 
drawn down to sill level. Safe working unit stress for steel members 
was assumed at 10 000 lb. per sq. in. 

The collection of drift in the gate recesses and in the gates, has 
given great trouble at Davis Island. This is due largely to the fact 
that that lock is filled partly by valves in the upper recess, and is 
emptied partly by valves in the lower recess. The operation of these 
valves creates currents toward the recesses, which draw into them the 
drift accumulated in the forebay and in the lock-chamber. It is not 
believed that as much difficulty will be encountered at the other locks, 
as they are filled and emptied entirely through valves in the river 
wall. However, in order to prevent any drift from entering the recesses, 
it has been proposed that the gates built hereafter be provided with 
pervious screens on their up-stream sides, placed so that when the gate 
is moved, these screens will pass close to similar screens projecting from 
the up-stream sides of the recesses at the entrances. All drift can 
then be locked through, and none will be allowed to enter the recesses. 

The gates at Locks Nos. 2 to 5 can be opened or closed in from 1^ to 
2 mln. The average time of operation at Lock No. 6, which has been 
in use for three seasons, is 2 min. The gates at Lock No. 6 have been 
moved while under a 6-in. head, the lower gate being closed while the 
filling valves were open and while there was considerable current in 
the lock-chamber. 

The recesses in which the gates rest when the dam is down, and 
from which they emerge when the lock is put into operation, are 125 ft. 
long and 20 ft. wide in the newest locks under consideration. The 
recesses of the lock nt Davis Island are so narrow that there is only 



THE IMl'lf()\ i;.MKNT OF THE OlflO RIVER -109 

about 1 ft. clearance on each side of the gate. This has often proved 
to be too small to permit of making the repairs to the gates, which 
are constantly required. On one occasion it was necessary to raise the 
gate 2 or 3 ft. off the track in order to make repairs. When a gate is 
erected in place on the tracks, which is the usual method, considerable 
room is needed during the erection. Moreover, if the gate be erected 
on the track in the recess after the lock-chamber has been flooded, a 
sump should be provided inside the entrance to the recess, and there 
should be room for the necessary suction pipes, etc. It is often desira- 
ble to use a horizontal, centrifugal pump in unwatering the recess, and 
if such a pump is not placed on some kind of floating support, it is 
desirable that it be placed in the recess. It is well, therefore, to make 
the recess large enough to permit this arrangement. It is believed that 
the recesses should be provided with vertical slots in the side-walls 
near the entrance so that horizontal timbers can be placed in the slots, 
thus forming the sides of a coffer-dam. 

The gate-recesses at Locks Nos. 2 to 6 are wider, but not longer, 
than those at Davis Island. It is believed that a recess should be about 
135 ft. long for a lock 110 ft. wide, the gate being about 118 ft. long. 
It has been noted that the gauge of the gate tracks is 11 ft. 6 in. 
They were so constructed, but, just prior to the erection of the gates, 
at Locks Nos. 2, 4, and 5, it was found that the gauge of the track 
at the entrance of the recesses was from 1 to 2 in. too small. Inasmuch 
as no greater clearance than | in. had been provided in the gauge of 
the gate tracks, it was necessary to tear up and reset a portion of the 
track. It is believed that expansion of the concrete in the long land 
walls and guide walls was responsible for the narrowing of the track. 
To prevent any movement of the track rails, it has been proposed to 
use ties made of two channels, back to back, with the rails riveted to 
an interior gusset plate at each end of the tie, these steel ties to be 
used instead of the former wooden ties. 

There having been considerable trouble at the Davis Island Lock, 
caased by the collection of drift in the gate recesses, drift chutes were 
built when the guide walls were rebuilt of concrete. In Locks Nos. 2 
to 6, the drift ch\ite is a tunnel running from the rear of the upper 
gate recess down past the rear of the lower gate recess (with which it is 
connected) and ernptying into the tail-bay of the lock. The entrances to 
the drift chutes are provided with movable doors, and are intended for 



410 



THE IMPROVEMENT OF THE OHIO RIVER 



flushing the drift out of the recesses. The construction of the drift 
chutes is very expensive, especially so when rock is found at a high 
elevation. It seems that it might be well to omit them and solve the 
drift problem by collecting the drift below the mouths of the tribu- 
taries, and not be bothered with the same drift at every dam on the 
river (Fig. 1). Of course, the greatest quantity of drift is running 
at times of freshets, when the dams are down, and the only drift to be 
cared for by the method proposed would be that which is floating on 
the pools. This will probably decrease in quantity. 



OHIO RIVER 

GENERAL PLAN OF LOCK AND DAM 




Fig. 1. 
The flushing conduits are small tunnels which run from near the rear 
of the gate recesses and empty into the lock-chamber and tail-bay. They 
are intended to be used in flushing all sedimentary deposit out of the 
recesses. They are provided with Stoney gate-valves at the entrances, 
and, if used when the corresponding gate is across the lock-chamber. 



THE IMPROVEMENT OF THE OHIO RIVER * 411 

a head equal to the lift of the dam is available for flushing purposes. 
They will work very nicely if the deposit is so small in quantity as to 
permit the gate to be run across the lock-chamber, but if the gate 
can be run across the lock-chamber, why is any flushing necessary? 
It is noteworthy that, although flushing conduits were built at Lock 
No. 6, they have never been used, and the lock has been operated con- 
tinuously for three seasons. After the high flood of March, 1907, the 
quantity of deposit in the gate recesses was very great. In some of 
them there was found a sedimentary deposit 7 ft. deep, and, from the 
two gate recesses at Lock No. 2, approximately 1 400 cu. yd. had to be 
removed. The deposit at all the locks after this flood was so great 
that, before any gate could be moved, it was necessary to coffer off 
the recess and remove the deposit. The flushing conduits were of no 
use because no head was available, the dams being down and the gates 
stuck fast. In order to ensure the satisfactory working of the lock- 
gates and to increase their durability, it is believed that it is wise to 
pump out the gate recesses every spring prior to the first raising of 
the dams. The debris can then be removed, the gates repaired, if 
necessary, and painted. 

Gate recesses are provided with roofs built either of reinforced 
concrete or of I-beams supporting ^-in. steel plates. The last-men- 
tioned type is believed to be the better, as it can be easily removed 
prior to making any needed repairs to the gates. The roofs are placed 
at such a height as will permit a man to walk on top of the gate and 
push the drift out of the recess, or into the drift chute. 

Each lock is provided with a conduit which runs across under the 
lock from the bottom of a well in the power-house to the bottom of 
another well in the river wall. It is in this conduit that the pipes 
are laid which supply pressure for working the river-wall valves and 
the bear-trap gates. In placing or repairing these pipes, it is neces- 
sary to pump out the conduits, and thoy should be provided with a 
sump at the bottom of one of the wells. If used at all, these power 
conduits should be placed so low that they cannot be injured by any 
dredging that may be required in cleaning out the lock-chamber. In 
the later designs, the conduits have been omitted, the power pipes 
being placed in troughs in the gate-track foundations. 

At Davis Island, the lock is filled through seven, circular, butter- 
fly valves, 4i in. in diameter, in the river wall, and the same number 



412 • THE IMPR0VE:\IEKT of the OHIO RI^'ER 

in the down-stream wall of the upper gate recess. The lock is emptied 
by seven valves of the same kind and size in the down-stream wall of 
the lower gate recess, which all discharge into a large conduit empty- 
ing into the tail-bay. Gate-valves were also used. The wall-valves 
were provided with vertical axes, and were turned by lever arms moved 
by a large hydraulic jack, one jack turning seven valves. The valve 
stems had rigid connections with the hydraulic jacks. Because of 
obstructions getting in the valves, most of the valve stems became 
twisted and the valves would not close, thus making the leakage great. 
A spring compensating device was originated, which permitted any 
one valve to remain open while the others were closed. At Lock No. 6, 
considerable trouble was had with the breaking of springs in this 
compensating device, and individual jacks, arranged in a parallel closed 
circuit, were adopted for use at Locks Nos. 2, 3, 4, and 5, each valve 
being provided with one of these little engines. The jack consists of a 
cylinder of Shelby drawn-steel tubing, 5 J in. in diameter, provided 
with cylinder heads, cross-head guides, etc. Motion is given to a lever 
arm by a connecting rod operated by a piston of 30-in. stroke. The 
parts were proportioned so that the valve shafts could not be twisted 
by any force which could be transmitted to the lever arm by the pis- 
tons. The shaft of any valve not operating would not be injured, there- 
fore, while the others turned. The valve shafts are vertical, and made 
of cold-rolled steel, 3| in. in diameter. 

The valves at all the locks, except those at Davis Island, were placed 
in the river wall of the lock — sixteen above the dam, for filling the lock- 
chamber, and sixteen below the dam, for emptying it. The valves are 
circular, butterfly valves, 4i in. in diameter, of cast iron, in a single 
piece, with horizontal stiffening ribs. 

The individual jacks have been highly successful. They are 
operated by water, except in the coldest weather, when oil is used. 
TIio pressure' is obtiiiiicd from ;i (lujilcx piiinp in tlie jjowcr-liouse. Tlie 
piping is installed so that, while water under pressure is supplied at one 
end of the jacks from the common supply line operating the pistons 
and valves, the water on the other side of the pistons flows into the 
other supply line and back to the tank in the power-house. A four- 
way valve controls the flow, so that the direction of motion may be 
reversed. 



PLATE XXIII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 




Fig. 1.— Manei-vering Boat at Lock No. 3, Ohio River. 




Fig. 2.— Opening the Dam at Lock No. 2. Ohio Rivkr. 



I 



iiiK i.\iri;(»\i:.\ii;.\i' ok ■riiio (tiiio i;i\i;i: 1 i;j 

Powp:r-Plants. 

At Davis Island, the steam for operating the lock-gate engines was 
generated in a boiler at the lower end of the locks; the steam for the 
upper gate engine was piped about 600 ft. to the engine; so much 
trouble was had from condensation that a boiler w^as installed at the 
upper recess. At Lock No. 6 the boilers were assembled in the power- 
house, which was opposite the middle of the lock, and steam was piped 
a distance of about 300 ft. to each gate engine; here again considerable 
trouble was encountered on account of condensation. A single-stage 
air compressor, with a capacity of 200 cu. ft. of free air per min,, had 
been installed to assist in moving the bear-trap gates. This was re- 
placed by a two-stage steam-driven compressor, with a capacity of 400 
ft. per min., and larger storage tanks were installed. Thereafter the 
gate engines were operated more satisfactorily with compressed air. 

The valve engines at Davis Island were originally operated by 
water pressure, supplied by elevated tanks under a head of about 62 ft. 
It proved to be hard to move the valves when under a head exceeding 
5 ft., and later the pressure was obtained directly from steam pumps. 

The plants at each of the recently completed locks consist of two 
Westinghouse, three-cylinder, vertical, gas engines, operating on 
natural gas, and two chain-driven, Blaisdell, two-stage air compressors, 
each having a free-air capacity of 500 cu. ft. per min., at a pressure 
of 100 lb. The operation of the bear-trap is facilitated by air forced 
into the lower leaves, direct from the air tanks. The gate engines are 
also operated by compressed air. The filling and emptying valves 
are operated by liquid pressure, as already stated, the pump being 
operated by compressed air. For maneuvering the gates, it is simply 
necessary to have a man at the throttle of the gate engines, but all 
operation of valves is from the power-house. Cooling water for the 
engines and compressors is obtained from elevated tanks supplied by 
an air-driven pump in the basement of the power-house. 

One great advantage in operating the valve engines by liquid 
pressure is that, in case of accident to the supply of gas, a small boiler 
can be coupled ([uic-kly to the supply pipe nt" the uprratiug jjuiup, and 
the valves can be operated in the same way as when the pressure pump 
was air-driven. If the valve engines were operated directly by air 
pressure, this could not be done. 



414 the improvement of the ohio river 

The Dam. 
The dams are principally of the Chanoine wicket type. The part 

of the dam over which navigation passes when the dam is down is 
called the navigable pass. In these dams, the navigable pass is located 
next to the lock, so that entrance to the lock will not be disturbed by 
currents from the weirs, etc. Each dam contains regulating weirs of 
the bear-trap variety. They are separated from the abutment shore 
by a weir of the Chanoine type, except at Lock No. 6, where there is a 
weir of the A-frame type. The current through the bear-traps is so 
swift and the scour below them so great, that, if they are placed next to the 
bank, considerable extra expense must be incurred for bank protection. 

In raising one of these dams, the pass is generally closed, first by 
the maneuvering boat; then the Chanoine weir is closed by a winch 
running on a service bridge, and the bear-trap, or automatic part of 
the dam, is closed last — the raising of the rest of the dam having 
created head enough to permit the raising of the bear-trap gates either 
by hydrostatic pressure alone, or assisted by air. (Plate XXIII.) 

The requisite weir area is important. It should be sufficient to per- 
mit of closing the pass without causing a rise of water surface which 
will prevent the maneuvering boat from raising the last wickets, the 
difficulty being in catching hold of the wicket. 

While the pass is being raised, discharge is taking place through 
the Chanoine weir, through the bear-traps, and through the part .of the 
pass not yet raised, as well as through the spaces between the wickets 
already raised. In calculations, all these points must be considered. 
Considering now the lowering of the dam because of a rise in the 
river, the discharge area should be manipulated so as to pass the 
stream flow without the necessity of lowering all the dams until 
the discharge becomes large enough to maintain a natural stage of 
9 ft. in the river. The upper pool may be allowed to rise until it is 
about 1 ft. above the crest of the dam, without iircvcnting the lowering 
of the dam from a maneuvering boat. 

It is important that the amount of automatic weir area be not made 
so great that the flow of the river through the openings can take place 
without creating head enough to operate the weirs. 

In designing the foundations of the dam, the height of the sills 
or highest fixed parts must be first decided. This is governed by the 
desirability of not interfering with the normal flow of the river and 



PLATE XXIV. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 



i <»,^«V 



OHIO RIVER 

CROSS-SECTION OF DAM N0.6 

DRAWN UNDER DIRECTION OF MAJOR W.H. HEUER, 

CORPS OF ENGR'S., U.S.A. 

JULY 10TH, 1896.' 




OHIO RIVER ■ 
CROSS-SECTION OF FOUNDATION OF DAVIS ISLAND DAM (DAM N0.1) 
DRAWN UNDER DIRECTION OF COL. W.E.' MERRILL, CORPS OF ENGR'S., U.S.A, 

DATE NOT KNOWN 




THE IMPROVEl^lKNT OK THE OHIO RIVER 415 

also by the requirements of navigation. The sill of the pass should be 
placed at least as far below low water as the crests of the nearest bars 
above and below the dam. If the sill is placed too low, there will be 
unnecessary trouble in handling the longer wickets, and the founda- 
tion will cost more because of being further below the surface of the 
water. If the sill of the pass is placed too high, it may interfere with 
open-river navigation. 

The sills of the weirs are placed at higher elevations than that of 
the pass. A Chanoine weir provided with a high sill will have short 
wickets, and can be easily operated so as to assist in passing small rises 
without lowering the entire dam, and thus lowering the pool. 

The various types of foundations used in these dams are shown on 
Plates XXIV, XXV, and Figs. 2 and 3. 

The general considorations which should govern in designing a 
dam foundation are, that there should be a water-tight surface at the 
up-stream face, and that the remainder of the dam should be con- 
structed so as to maintain the water-seal in place and prevent its 
disturbance by scour below the dam or tmderneath it. In all these 
dams the water-seal is practically the same. The dams differ in the 
style of support for the concrete, and in the character of the protection 
against scour below the dam. Below Dam No. 2, and below portions of 
Dams Nos. 3, 4, and 5, stone-filled apron cribs were used, heavy rip-rap 
being placed down stream from the cribs. Below Dam No. 6, and 
below part of Dam No. 4, protection against scour is afforded by a 
wide apron of rip-rap held in place by piles driven in quincunx order. 
This style is similar to that used in the dams on the River Yonne. 
Below a portion of Dam No. 5, a wide concrete apron, in which were 
embedded old wire ropes, was used. 

The question of scour below dams is of great importance. It has 
been necessary at Davis Island on many occasions to place rip-rap 
below the dam, at one time whole barges loaded with stone being sunk 
as a protection. The scour is especially bad both above and below 
the bear-trap gates. Soon after Dam No. 6 was put into operation, 
scour developed to such an extent around the piers and bear-traps 
that it reached to the bottom of the concrete at the head of the piers, 
and just below the protection apron, the scour went to bed-rock, about 
31 ft. below low water. It was necessary to place about 2 000 tons of 
rip-rap to stop this scour. 



416 



THE IMl'HOVEMENT OF THE OHIO RIVEIl 



CourBce uf 1 plank 








33 

m 


o 




f) 


AJ 




H 


o 




U 


CO 




^ 


CO 




o 

■n 


C/5 

m 


O 


> 


O 


X 




H 


— 


H 


O 


O 


? 


7' 




?• 




JJ 


p 


o 

-n 


< 

m 


03 


O 


;a 



UD CO 



Tllli IMrUOVElSrENT OF TIIK OHIO HIVKR 



417 




418 THE IMPROVEMENT OF THE OHIO RIVER 

The type of foundation shown in Fig. 3 is that used, in the latest 
foundation work on the river, when rock at practicable elevations does 
not exist. The water-seal is formed by the wickets, by the up-stream 
edge of the concrete, and by a row of 9 by 12-in. Wakefield sheet-piles. 
The concrete is supported on round piles, driven to rock if possible. It 
is a general principle of construction that the upward pressure beneath 
a dam should be made as small as possible. This is done by making 
sure that the up-stream face of the dam is more nearly water-tight 
than the down-stream face. As there are bound to be leaks between the 
sheet-piles, an effort has been made in designing these dams to provide 
free egress for the leakage water in such a way as to insure no move- 
ment of the material under the dam. This is done by passageways 
leading up from underneath the tail of the concrete and opening into 
the cribs. In order to permit of inspection of the condition of the 
foundation beneath the dam, vertical holes, 6 in. in diameter, were left 
in the concrete near the down-stream edge, at intervals of 8 ft. from 
center to center. These were plugged with swelled wood. The plugs 
can be removed, or holes can be bored through them and soundings 
taken to ascertain the conditions existing underneath the dam. In case 
of scour, the opening can be filled with grout or concrete. 

At some of the dams, a wooden covering was placed on top of the 
foundation concrete, the timbers being anchored to the concrete. The 
idea was to prevent damage to wickets by interposing an elastic cushion 
between them and the concrete. It will soon be necessary to repair the 
wooden top at Davis Island, and it will be a hard task. The wear 
at the top has been caused very largely by the props running out of 
the hurters and gouging holes in the wood. It might be advisable in 
designing the foundation for a movable dam of the Chanoine type to 
arrange the top surface so as to fit a movable coffer-dam which could 
be used in replacing movable parts, making minor repairs to the 
foundation, etc. 

The wickets used in the Chanoine dams on the Ohio Hiver are 
3 ft. 9 in. wide, and the largest are approximately 18 ft. long. As they 
are placed 4 ft. apart from center to center, there are 3-in. spaces be- 
tween the wickets, which serve to prevent interference of the wickets, 
due to an oblique pull in raising, etc. The discharge of the Ohio River 
is sufficient to permit of this arrangement, and when the discharge is 
very small, , wooden needles or joint covers are used to cover the spaces 



PLATE XXV. 

TRANS. AM. SOC. CIV. ENGRS 

VOL. LXIll, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 



OHIO RIVER 
CROSS-SECTION OF DAM NO. 4 ADJACENT TO LOCK. 

DRAWN UNDER DIRECTION OF MAJOR W.H. BIXBY, CORPS OF ENGR'S., U.S.A. 

APRIL 6Tli, 1901. 




^I^J^jg^:^^^£^Sgj!v^a^a^^.;;..J>^'^/.»^>;jfev^^^^^ 




■l«-'0- 




-25'0- 



C0M.C5ETE 






■ ;;_:;;,■ :^-r';~gy{a^^ -;.:...;..li;y.:y ^ !.. /.^ .. ■ " V 




OHIO RIVER 

CROSS-SECTION OF FOUNDATION OF DAM NO. 5 ADJACENT TO LOCK, 

DRAWN UNDER DIRECTION OF CAPT.W.E. CRAIGHILL, CORPS OF ENGR'S., U.S.A. 

JUNE, 1902. 



THE IMPROVEMENT OE THE OHIO RIVER 419 

between the wickets. The wearing away of the wickets lias sometimes 
made it impossible to stop all leakage at the Davis Island Dam, even 
when the needles are used, and, at times, straw and willows have been 
used to decrease the leakage. This pool, however, has been maintained 
for years with no pool below, when leakage was greater than it will be 
hereafter, this part of the system having been completed. 

The pass wickets should be designed so that they will not swing, at 
least until the \ii)per pool rises to a lieiglit of 2 ft. above the crest of the 
dam. The water has been allowed to rise that high at the Chanoine 
Dam in the Allegheny Eiver; however, that was under exceptional 
circumstances. It is generally necessary to commence to lower the 
pass by the time the upper pool is 1 ft. above" its normal level. In 
order to assist the wicket in maintaining its upright position, cast-iron 
counterweights are placed in the breach of the wicket. The desira- 
bility of making the pass wickets non-automatic until a certain stage 
is reached, puts an inferior limit on the distance from the foot of the 
wicket to the journal boxes, and the economy of keeping the length of 
the horse and prop small provides the superior limit. 

It is customary, in the operation of a Chanoine dam in the Ohio, to 
leave one wicket of the pass down near the river wall of the lock. 
This affords a passage for small drift which collects at times near the 
lock wall. 

The horses have a large factor of safety when the normal tension 
to which they are subjected is considered; but very often they have 
been damaged by being struck by boats, snags, etc. Those in the dams 
described generally have been provided with side bars IJ by 3 in. in 
section. The latest ones have 2 by 4-in. side bars. At first, the axles 
were made of cast steel, but many were broken, and forged steel axles 
were then adopted. 

The props have been 3i to 4 in. in diameter, with a large excess of 
material at the lower end, in order to ensure that they will remain in 
position in the hurter channel when the dam is being raised. The eddy 
currents have a tendency to swing the props sidewise, and to prevent 
their seating properly against the step in the hurter. It has been 
considered by some that the large bulge at the bottom end of the prop 
contributes to this by offering a greater surface for the water to act 
against, and of late experiments have been made with cylindrical props, 
5i in. in diameter throughout their length. It sometimes happens in 



420 TILK IMPROVEMENT OF TJIE OHIO RIVER 

lowering tlie wickets that the props do not run back properly on the 
hurter grooves, but catch and prevent the wicket from falling. It is 
then customary to lift the lower end of the prop with a hook on the 
end of a line from the maneuvering-boat derrick, this hook engaging 
in a notch in the lower part of the prop. 

The current increases greatly ns the last few wickets are reached 
in raising the dam, and these wickets should be provided with heavier 
props than the other wickets of the pass. One of the greatest troubles 
encountered in raising the pass is the difficulty of catching hold of 
the handle plates of the last few wickets to be raised. It is believed 
that the wickets of the pass nearest to the first pier might well be pro- 
vided with chains reaching from the bottom handle plate of one wicket 
to the top handle plate of the next wicket on the side toward the lock. 
The last few wickets can then be raised by these chains. 

The Bear-Traps. 

A bear-trap weir was installed at Davis Island in 1889. The gates 
were of wood, and were 52 ft. long, forming a weir of the type known 
as the "old bear-trap." It was intended for use as an automatic weir 
for the regulation of the pool levels, and also as a drift-pass. This 
bear-trap weir proved to be of much assistance in operating the dam. 
After the original bear-trap at Davis Island had been in service for 
about sixteen years, it was replaced by a new bear-trap, the leaves being 
of wood heavily bound with steel, and proportioned on the basis of 100 
upward lifting force to 66 downward. 

These new leaves are much stronger and more rigid than the old 
ones, and work very satisfactorily. The thought in the design was to 
build the leaves so that they would just about float, and thus elimi- 
nate the need of air. For lengths not exceeding 60 ft., the plan is very 
promising. 

At Dam No. 6 two bear-traps were installed, each being 120 ft. long. 
The principal members were of steel filled in with wood. The leaves 
were proportioned on the basis of 100 lifting force to 80 downward. In 
order to assist in the raising of the gates, arrangements were made for 
increasing the buoyancy by air forced to the under side of the lower 
leaves. The air was intended to be pocketed between the ribs of the 
girders. This proved to be of little value, because the air could not be 
retained where it was needed. If both bear-traps are up and one is 



PLATE XXVI. 

TRANS. AM. SOC. CIV. ENQR3. 

VOL. LXIII, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 




Fig. 1.— Beak-Trap Dam. Lock No. 5, Ohio Riveh. 




Fig. 2.— Bear-Trap Dam, Lock No. 5, Ohio River. 



THE IMPROVKAfEXT OF THE OHIO RIVER 421 

lowered temporarily, air is not used in raising it again. It takes a 
head of aliout 2A to 3 ft. to raise the ^ates if air is wsi'd, and al)out 
3i ft. without it. These gates rise in from 6 to 8 min., if under a 
9-ft. head, and can be lowered in from 3 to 4 min. So much of this 
dam is of the automatic type — about 25% of the entire dam — that it 
is difScult to secure the head needed to raise the gates if both are 
down. 

Each of the lower leaves of the bear-traps in Dams Nos. 3, 4, and 5, 
consists of 19 girders, spaced 5 ft. apart, center to center. There are 
also five longitudinal braces, called girts, which divide the leaf into 72 
sections. Certain girders are solid throughout, and divide the leaf 
into three water-tight compartments. All the girders are hinged, at 
the down-stream end, to steel castings anchored to the foundation. The 
up-stream ends are provided with rollers which support the girders of 
the upper leaves as they slide over the upper end of the lower leaves. 
Each of the steel castings, which holds the rollers, terminates in a 
double hook which engages in stops on the upper end of the up-stream 
leaf, when both leaves reach the limiting height. In order to hold the 
gates in the upright position during repairs, etc., locking-pin holes are 
provided. The skin-plating of the lower leaf consists of buckle plates 
and flat plates. Manholes are left in the plates, those in the upper side 
of the leaf, being provided with covers, are intended for use only in 
cleaning the leaf, making repairs, etc., while those in the lower side 
are intended to permit the water in the leaf to be displaced by air. The 
lower leaf when filled with air is calculated to have such buoyancy as 
will cause the entire bear-trap to rise in still water, no hydraulic 
lifting force being needed. The air enters the pockets of the lower leaf 
through special ball and socket pipe joints. 

Leakage from beneath the gates, between them and the piers, is 
prevented as far as possible by water-seal arrangements placed so as to 
be pressed against the sides of the piers. At the lower end of the 
gates, curved plates attached to the leaves move close to similar plates 
attached to the hinge castings. (Plates XXVT and XXVH.) 

The question of leakage prevention is of great importance, and 
should be carefully considered. The bear-trap gates at the Herr Island 
Dam in the Allegheny River could not be raised automatically until the 
space between the upper and lower leaves was decreased by thin boards. 
They can now be raised under a 4i-ft. head witliout the use of air. 



422 THE IMPROVEMENT OF THE OHIO RIVER 

The upper leaves of the bear-traps at Dams Nos. 3, 4, and 5, consist 
of a series of nineteen I-beams, spaced 5 ft, apart, center to center. 
These girders are separated by longitudinal braces. Each I-beam is 
hinged at the up-stream end to a steel casting anchored to the founda- 
tion. The leaf is sheathed with dressed white oak, carefully fitted 
between the I-beams. This leaf does not have to support any water 
pressure, and is made as described so as to decrease its weight. 

The water for operating these gates is taken from the sides of 
Piers Nos. 1 and 3 farthest from the bear-traps. This is because there 
is so much fall around the heads of these piers, and it is important to 
use all the upward pressure that can be obtained. The conduits are 
built so that the water enters each bear-trap from each end. This 
was done in order to decrease the tendency of the leaves to warp. These 
traps are proportioned on the basis of 100 upward lifting force to 60 
downward. They have not yet had sufficient trial to permit of the ex- 
pression of opinion as to their merits. The upper end of the foundation 
is arranged for supporting a needle-dam, and this can be used to secure 
the head needed to raise the gates, if impossible to raise them other- 
wise, and can serve as a means of holding the pool while they p'*'^ 
being repaired. 

Chanoine Weirs. 

The Chanoine weirs, which are located between the bear-traps and 
the abutments of all these dams except Dam No. 6, are similar to the 
passes. The wickets are raised, however, by a winch running on a 
structural-steel service-bridge, standing on the weir foundation above 
the wickets. 

It was originally intended that the entire dam at Davis Island 
should be operated from a service-bridge, but before its completion it 
was decided to place the journal boxes for the bridge in the pass foun- 
dations, but not to erect the bridge except above the three weirs. The 
pass was operated from a small maneuvering boat. During a sudden 
freshet in 1887, the bridge of Weir No. 1 was destroyed by drift, and, 
in 1889, the bridge of Weir No. 2 was destroyed in the same way. On 
a later occasion, the bridge above Weir No. 3 was struck by a coal barge 
during a freshet, and partially wrecked. 

The original trestles were composed of rather light members, and 
those built later have been made much stronger. It seems to be ad- 



PLATE XXVII. 

TRANS. AM. SOC. CIV. ENQRS. 

VOL. LXIIl, No. 1104. 

SIBERT ON 

IMPROVEMENT OF THE OHIO RIVER. 




Fig. 1.— Bear-Trap Dam, Luck No. n, Ohio River. 




1^ 







Fici. 2.— Lock Xo. 1. Alleghe.n'Y River. 




I'l,;. :!. — !. (.IK X". 1. I '111(1 l\ivn!. 



THE IMPROVEMENT OF THE OHIO RIVER 423 

visable, however, to avoid the use of service-bridges in the navigable 
passes of the Ohio Kiver for the following reasons: 

^■^^•""^^^ bridge and its foundations constitute an extra expense. 
(2).— Likelihood of destruction of the bridge by drift or ice. 
(3).— The bridge retards the lowering of the dam, because, after 
the wickets are lowered, it is still necessary to lower the bridge. The 
lowering of the pass can be effected quickly by a maneuvering boat. 

The latest weir service-bridges for these dams have trestles 8 ft. 
apart, center to center, the trestles being arranged to rotate about their 
bottom axles, which turn in journal boxes anchored to the foundation. 
The down-stream journal box is open on top, the axle being prevented 
from coming out by an iron key. When raised, the trestles are kept 
in place by steel floor-panels, each hinged at one end to the top of a 
trestle, and held at the other end by pins attached to the adjoining 
trestle, the pins passing through holes in castings attached to the end 
of the floor-apron. The floors form a track for the winch used to 
operate the wickets, and serve also as a footbridge for the dam-tenders. 
An effort was made to make these bridges strong enough to resist 
the destructive effect of drift, and they were also made strong enough 
to support the full upper-pool pressure, assumed to be borne by needles 
resting at the bottom against the oak sill in which the upper journal 
boxes are embedded, and at the top against horizontal beams resting 
against the tops of the trestles. 

A rolling, hand-operated winch has always been used to maneuver 
the wickets at Davis Island Dam. The winch is provided with a drum 
parallel to the crest of the dam. Another winch having a drum parallel 
to the direction of the current, is used to raise and lower the bridge 
trestles successively. 

The operating winches for Dams Nos. 2 and 5 are to be operated by 
compressed air, furnished from the power-house through a 4-in. pipe, 
embedded in the foundation of the dam and ending at the abutment. 

The maneuvering boats used to operate the wickets of the passes 
are provided with derricks, hoisting engines, steam capstans, etc. The 
hull of one of the latest of these boats is 70 ft. long, by 22 ft. wide, 
and 3 ft. deep, and is built of wood throughout. A stiff-legged derrick 
was placed near the bow of the boat; this is operated by a three-drum 
hoisting engine. 

The wickets are raised by a wire line leading from the hoisting 



424 THE lAIPROVEMENT OF THE OHIO RIVER 

engine around a sheave held at the end of a structural-steel beam, 
which projects from the bow of the boat. The wire line is fastened to 
an iron hook at the end of a long pole, which is used to guide the hook 
until it engages in the handle of the wicket. After the prop is seated 
in the hurter, the wicket is assisted in seating itself against the sill 
by pike poles in the hands of the dam-tenders. In lowering the wickets, 
the upper ends are pulled slightly up stream, until the prop unseats itself, 
when the wicket is allowed to fall upon the foundation behind the sill. 

The boat is prevented from touching the wickets above their axes 
of rotation by spuds which bear against the wickets below the tops of 
the horses. The current around the end of the wickets already raised 
is very swift. A manila rope, of 2 in. diameter, is fastened at one end to 
the river wall of the lock, and, at the other end, it is wound around the 
steam capstan on the stern of the boat. This rope prevents the boat 
from being swept around the ends of the wickets, and also serves to 
draw the boat back to the river wall after all the wickets have been 
raised. 

The abutments of these dams are not subjected to the severe con- 
ditions which exist at the usual fixed dams, because the lifts are smaller ; 
and when a freshet comes, the dams are lowered, thus restoring approxi- 
mately the normal discharge area of a stream. However, it is necessary 
to design the abutments so that they will be free from danger of 
failure, due especially to eddies produced by the flow of water through 
the bear-traps and by the current of water flowing through the weirs. 
If the abutments are not founded on rock, they rest on bearing piles, 
and are protected, not only by a row of Wakefield sheet-piling, but also, 
below the dam, by stone-filled cribs placed in front of the sheet-piling. 
The abutments are of the U-type, with a wall extending down stream, 
the function of which is to prevent the washing away of the bank below 
the abutment. It is deemed necessary to carry this wall for a distance 
of from 100 to 200 ft. below the crest of the dam, depending on the 
character of bank material, value of property, proximity of buildings, 
etc. 

The operation of such a system of movable dams as that described, 
demands a very careful watch of the discharge of the streams upon the 
head-waters. This is provided for by reports received daily, or more 
often if necessary, from observers stationed at various points. It will 
probably be necessary to construct a Government telephone or telegraph 



■Illi: IMI'lJdVEMENT or TJIE UliiO UIVKU 425 

line along the river, as has been done along the canalized Great 
Kanawha. Any increase in the low-water flow beyond that necessary to 
furnish water needed for (■\;i|)(iration and npcration, adds to the diffi- 
culties of the operation of such a system. 

In operating- a sinjile dam, there are many thinjis to be c(nitciidi'(l 
with. At Davis Island, jjoats have been known to hit a pier, a licar-trap 
gate, and lock-gates; there have also been many collisions with wickets. 
The fact that a boating coal barge destroyed a portion of the service- 
bridge has bfcu ut)te(l. However, if enough spare wickets, horses, and 
props are kept on hand, tliost- injured can be replaced by a diver without 
much difficulty. Some of the movable parts were broken by the stern 
wheels of steamboats before the dams were in operation, and it is very 
necessary to provide a sufficient number of spare parts. Notwith- 
standing all difficulties, Davis Island Dam has been operated success- 
fully for twenty-two years, and Dam No. 6 for three years. 

The 150 pass wickets of Dam No. G have been raised in 1 hour 40 
min., the river being at a natural stage of 6.4 ft., and the dam has 
freqiuMitly been raised in 2i hours. This dam can easily be lowered in 
1 hour. (Fig. 3, Plate XXVII.) 



426 DISCUSSION ON THE IMPROVEMENT OF THE OHIO RIVER 

DISCUSSION 



Mr. Ripley. Theron M. Eipley, Assoc. M. Am. Soc. C. E. (by letter) . — A read- 
ing of Major Sibert's paper brings again the query which has come to 
mind many times in the past few years, viz., on what data and after 
how careful consideration of the questions has been based the assump- 
tion that a possible 11 ft. is the maximum which should be provided 
for on the Ohio improvement, and if 9 ft. is economically necessary 
at Pittsburg at present, would not 11 ft. or more be the economical 
development below Portsmouth or Cincinnati? 

This query is not an insinuation as to the paucity of data or lack 
of study of the scheme for the Ohio River as a river, but in its relation 
to the possibilities and probabilities of contiguous improvements and 
their bearing on the Ohio River work. 

The State of Ohio contains at least one, and maybe two, routes 
along which it is possible to construct a canal from Lake Erie to the 
Ohio River with a depth of water of not less than 12 ft. 

For several years a determined effort has been made (and is now 
being made) by some of the State's best men to have such a canal con- 
structed. There are those who believe that the State could expend no 
money which would be of greater economic benefit than in building 
such a canal. 

Ohio is geographically situated directly between the immense ore 
deposits of Northern Michigan and the no less immense deposits of coal 
in West Virginia. Already traversed by the trunk lines of nearly all the 
railroads from the Atlantic to St. Louis and northern points, what more 
natural than that some of her citizens should believe in bringing this 
coal and iron together by the cheapest method, and shipping her manu- 
factured product by her own waterway and the Barge Canal of New 
York State to New York City, or by her railroads to Atlantic and inland 
points farther south, or by her canal and river to New Orleans and 
intermediate points? 

Questions such as these should be taken into consideration in any 
development for the Ohio, as any structures in that river will determine 
the economic navigable depth of connecting waterways above, and may 
assist or destroy their usefulness. In fact, a less depth below Ports- 
mouth or Cincinnati than that possible across the State of Ohio might 
prevent the building of an Ohio Canal, and in any event would be a 
serious handicap thereto. 
Mr. sibert. WiLLiAM L. SiBERT, M. Am. Soc. C. E. (by letter). — The following- 
query by Mr. Ripley is very pertinent: 

"On what data and after how careful consideration of the questions 
has been based the assumption that a possible 11 ft. is the maximum 
which should be provided for on the Ohio improvement, and if 9 ft. 
is economically necessary at Pittsburg at present, would not 11 ft. or 
more be the economical development below Portsmouth or Cincinnati?" 



DISCUSSION ON THE IMPROVEMENT OF THE OHIO RIVER 427 

Tlie Ohio and Mississippi Rivers were considered as one system in Mr. sibert. 
the project for the Ohio River. At present a low-water depth of 9 ft. 
is maintained between Cairo and New Orleans by hydraulic dredges. 
With an actual depth of 9 ft. at low water, a fleet of towboats and 
barges drawing more than 6 to 6i ft., would not attempt to navigate 
the Mississippi River. This is due to the narrowness of the channels 
at the crossings, and to danger from logs and snags embedded in the 
bottom and banks of this stream. The draft of single boats would 
approach more closely to 9 ft. 

The practicability of further improvement of the ]\Iississippi River 
is now under consideration. There are those, whose experience causes 
their opinion to be of value, who think that it is practicable to main- 
tain an open-channel depth of 14 ft. at low water in the Mississippi 
between Cairo and New Orleans by dredging and bank protection. 

A slack-water depth of 11 ft. in the Ohio River would practically 
provide for towboat navigation of the same draft as would a 14-ft. 
open-channel depth at low water in the Mississippi River. In a slack- 
water system there are no currents to contend with, and the limiting 
depths and contracted channel widths exist only in the heads of the 
pools, both of which conditions disappear after leaving the locks. 

The 9-ft. depth in the approved project for the Ohio corresponds 
therefore to the low-water depth at present maintained in the Missis- 
sippi for packets, and to a 12-ft. low-water depth for towboats. The 
Mississippi is naturally at or above a 12-ft. stage from Cairo to New 
Orleans for a large portion of the year. 

Considering the Ohio and Mississippi Rivers then as a trunk line 
of a great water-transportation system, these would be the reasons for 
the projected maximum depths in the Ohio River part of it, if such 
maximum depths were not fixed by other considerations. In the Ohio 
itself such considerations are cost, hindrances to navigation by many 
locks, and the practicable height of movable dams across great streams. 

The maximum attainable depth on the tributary streams is also 
a determining factor as to depth in the trunk-line stream. Such 
tributary streams constitute the feeder lines, and they should be of the 
same gauge, or the same navigable depth, or, at least, should be of 
such depth as to permit the transportation of material from such 
feeder to its market economically, and without transferring the cargo. 

Many of these tributaries can only be properly improved with 
locks and fixed dams, their steep slopes making open-channel navigation 
impracticable, thus excluding movable dams, were they not otherwise 
excluded by the certainty of sudden floods accompanied by ice or drift. 

Nearly all these streams transport considerable silt, and any slack- 
water plan for the improvement of silt-bearing streams which involves 
deep dredging in the upper ends of the pools must take into con- 
sideration excessive dredging in maintenance. Fixed dams being neces- 



428 DISCUSSION ON THE IMPROVEMENT OF THE OHIO RIVER 

Mr. sibert. sary, and dredging eliminated, it is seen that such dams must be 
high or very close together, if depths of as much as 9 ft. are to 
be made in the upper ends of pools in the tributary streams of the Ohio. 
Increase of flood damage soon limits the height of fixed dams, leaving 
out of consideration the fact that such increase decreases the flood 
velocity of currents and may cause the silting up of pools. Fixed 
dams, however, can be built higher than one vpould think, without 
investigation, and not materially change flood heights. A few inches 
fall at the dam in extreme flood so increases the discharging capacity 
as to enable such flood to pass the contracted section with only the few 
inches increase of height near the dam. 

As to assembling material for manufacture, such as coal and iron 
ore, on the Ohio River, and transporting the iron ore from the Great 
Lakes via a canal and the coal via the Ohio River and some of its 
tributaries, the maximum practically attainable depth in the tributary 
from which the coal comes would probably fix the maximum needed 
depth in the Ohio, as far as coal is concerned. Coal in fleets can be 
transported on the Ohio on a 9-ft. slack-water depth at a cost of about 
2 mill per ton-mile. 

Increase of cost of transportation, as depth decreases, is more rapid 
in contracted channels, such as canals, than in open rivers, where 
increased cargo can be obtained by covering more area with the fleet. 
This leads to the general conclusion that where a large freight move- 
ment is certain the depth in a canal should be governed largely by the 
available water supply. Should a canal be built connecting the Great 
Lakes with the Ohio River, the river could be improved to the canal 
depth, if practicable, for such a distance as to reach a suitable place 
for the assembling of material for manufacture and the distribution 
by rail of finished products. The shipment of the manufactured 
I)roducts to New Orleans would be by towboat, with barges of such 
draft as the improved Mississippi would fix, and which, it is thought, 
would not exceed the maximum stated for the Ohio. 

A depth of 11 ft. on miter sills has been adopted, it is understood, 
in the Barge Canal of New York. This depth can be obtained in 
the Ohio, under the 9-ft. project, by a small amount of dredging in 
the upper ends of the pools, the depth of water on the miter sills 
being 11 ft. 

Unless there were sjx'cial reasons for a greater depth than about 
12 ft. in a canal from the Great Lakes to the Ohio — such as the 
proximity of suitable and extensive manufacturing sites to the Ohio 
River terminus of siu*h a canal — tlie depth of water in the Barge Canal 
of New York and the nuixinuun obtainable towboat draft in tlic Ohi^^o- 
Mississipi)i would be strong arguments for a standard depth oi 1 1 ft. 
on miter sills in a canal (•(Hiiieet iiig the (Jrcnit l^akes and the Ohio. 



MEMOIR OF WILLIAM Bi:VEHLY CHASE 429 

MEMOIKS OF DECEASED MEMBERS. 



WILLIAM BEVERLY CHASE, M. Am. Soc. C. E. 



Died October 26th, 1908. 



William Beverly Chase, the son of Levi W. and Harriet Vining 
Chase, was born in Marengo, Morrow County, Ohio, on November 21st, 
1852. His family was of Colonial ancestry, his great-grandfather, 
Beverly Chase, having served as a New York Militiaman in the War 
of the Revolution. 

His father was an architect and master builder of the old school, 
and, in his early years, the son lived in the atmosphere of the builder 
and the artist and became acquainted with the use of the drafts- 
man's tools. 

His early education was secured in the high schools provided in 
the Central Ohio towns of that period, and has been described as 
''consisting mainly in digging out things for himself," and, to his praise 
be it said, that during his after life he made most successful use of the 
ability thus acquired. 

In early manhood he removed with his parents to Southwestern 
Minnesota, where for five years he was engaged in surveys and in 
acquiring practical experience in the design and construction of rail- 
road bridges. 

In 1877 Mr. Chase removed to Oregon, and for more than thirty 
years was identified with the growth and development of that State. 
For the first three years he was occupied with map work, and was 
in the employ of local bridge builders, but, in 1880, his opportunity 
came with the construction of the Northern Pacific Railroad, surveys 
for which were then in progress. Beginning as a Topographer for a 
field party, he was soon at Headquarters in Portland preparing designs 
and plans for the large number of structures required for the Western 
Divisions of the railway, then under the charge of V. G. Bogue, 
M. Am. Soc. C. E. 

From 1884 to 1885, he was Engineer of Bridges for the Oregon 
Pacific Railroad, a line crossing from the Coast through the Western 
portion of the State, of which the late Isaac W. Smith, M. Am. Soc. 
C. E., was Chief Engineer. From 1885 to 1890 he was engaged in 
hydraulic and sanitary surveys and work and in bridge construction 
for some of the towns of the State, among which were Corvallis and 
Eugene. 

In 1891 Mr. Chase made siirveys and designs for a sewer system 
for East Portland, and, from 1891 to 1895, he was Engineer of Bridges 
for the Portland Bridge Commission, during which time he con- 
* Memoir prepared by D. D. Clarke, M. Am. Soc. C. E. 



430 MEMOIR OF WILLIAM BEVERLY CHASE 

structed the Burnsido Street Bridge, crossing the WiUamette River, 
costing $300 000, and a steam ferry-boat for North Portland. From 1894 
to 1896 he was engaged in general practice, while from 1896 to 1902 
he was in the service of the City, first as Superintendent of Streets 
and later as City Engineer, retaining the latter position imtil, by a 
change in political parties, the wing with which he had allied himself 
suffered defeat. 

During all these years and until his last illness, he continued to 
act as Consulting Engineer for the County Commissioners, who under 
the law are charged with the duty of maintaining all the bridges and 
ferries crossing the Willamette within the City limits. After the 
retirement from the service of the City, he was engaged in making 
surveys and designs for various towns for water supply, street pave- 
ments, etc., the Towns of Astoria, Corvallis, McMinnville, Eainier, 
and Tillamook, Oregon, and North Yakima, Washington, being among 
the places which he served acceptably. 

It was while engaged in making an examination of a water-supply 
project for the Town of McMinnville, in July, 1908, that Mr. Chase 
suffered a severe paralytic stroke from which he never recovered, being 
confined to his bed from that time until his death, which occurred 
October 26th, 1908, at the Good Samaritan Hospital, Portland. 

Mr. Chase was known as a genial gentleman and a man of recog- 
nized ability and worth. He was a Christian, from early life, and 
although not demonstrative he was ever known as a loyal supporter 
of all things that make for that "righteousness which exalteth the 
nation." He was long connected with the Centenary Methodist Episcopal 
Church of Portland, Oregon, and by his loyalty and steadfastness was 
largely instrumental in sustaining it during a critical period of its 
history. His zeal in the service of his church was his by inheritance 
from a godly ancestry. 

In 1884 Mr. Chase was married to Miss Georgia Parker of Astoria, 
Oregon. The death of his wife in 1894, leaving him with their 
family of three little daughters, and the care of his aged father, greatly 
modified Mr. Chase's purposes and efforts in his profession, obliging 
him to put aside congenial opportunities offering exercise of his 
energies in wider fields, but through all his life he was an inspiring 
example of what may be accomplished alone, following a natural bent, 
supplemented by faithful application and courage. 

Mr. Chase is survived by two daughters. Misses Marion and Jessie 
Chase, of Portland, Oregon, and by a brother, the Rev. Charles E. 
Chase, of San Francisco, and a sister, Mrs. Lucia C. Bell, of Fruitvale, 
California. 

Mr. Chase was elected a Member of the American Society of Civil 
Engineers on September 6th, 1899. 



MEMOIR OF MARTIN WILLIAM MANSFIELD 431 

MARTIN WILLIAM MAiNSFIELI), M. Am. Soc. C. E.* 



Died September 25tii^ 1908. 



Martin William Mansfield was born at Ashland, Ohio, on November 
19th, 1S50. His father, Martin H. Mansfield, was of English birth; 
his mother was Anna Saiger, of Mifflin, Pennsylvania. 

Mr. Mansfield was graduated from Rensselaer Polytechnic Institute 
with the Class of 1871. In September of the same year, he entered 
the service of the Pennsylvania Lines West of Pittsburg as Assistant 
Engineer in the Maintenance-of-Way Department on the Cincinnati 
and Muskingum Valley Railroad (a subsidiary line), at Zanesville Ohio. 
He was promoted successively to Engineer of Maintenance of Way, 
Superintendent, and Assistant Chief Engineer, which position he held 
at the time of his death. 

On June 24th, 1878, Mr. Mansfield married Miss Carrie Sampsell 
at Ashland, Ohio, who survives him, with their son, Sampsell W. Mans- 
field, and daughter. Miss Corinne S. Mansfield. 

As a student at Troy Mr. Mansfield was diligent, earnest, and suc- 
cessful. One of his classmates writes of him, "he gave evidence at that 
time of the unusual talent, that crowned his later years, for working 
out difficult and abstruse mathematical problems." This talent was 
indeed characteristic of the man, and was frequently called into play 
by special lines of investigation assigned to him by his superior officers, 
who recognized his ability to analyze a mass of apparently heterogene- 
ous facts, reduce them to order, and find the underlying fundamental 
principle. 

Kindly, afi'able, and accessible, but strict in discipline, he com- 
manded the esteem and good-will of his subordinates. Earnest, con- 
scientious, and upright, in all things, he had the confidence of his 
superiors. Quiet and unassuming in manner, cheerful in disposition, 
and equable in temper, he won the respect of all who came in contact 
with him. 

Mr. Mansfield was elected a Member of the American Society of 
Civil Engineers on July 5th, 1882, and by his death the Society loses 
one whose professional abilities and private character were an honor 
to it. 

* Memoii- prepared by Thomas H. Johnson, M. Am. Soc. C. E. 



433 MEMOIR OF MARK WILLIAM SCHOFIELD 

MARK WILLIAM SCHOFIELD, M. Am. Soc. C. E.^ 



Died November 27tii, 1908. 



Mark William Schofield was born in Smithfield, lihode Island, on 
November 10th, 1846. His parents were of English stock, and came 
to the United States about the year 1844. His father died while he 
was very young, and his mother some years later, in 1864. During 
his early life he attended the village school in Georgiaville, and later, 
the Lapham Institute, in Scituate, an advanced academy where many 
men prominent in later life received a large part of their education. 

While attending school his energetic nature led him to spend a 
portion of his time in the mill in his native town, where he acquired 
much practical information regarding the details of cotton machinery. 
His tastes, however, led him toward the profession which he afterward 
pursued, and early in 1867, and previous to July, 1868, he was for a 
time with Mr. William S. Haines, and with Gushing and DeWitt, two 
of the older surveying and engineering firms of Providence. In. July, 
1868, Mr. Schofield went West, and was engaged on the surveys of 
the Cairo and Vincennes road, with which General Ambrose E. Burn- 
side was at that time closely identified. Desmond FitzGerald, Past- 
President, Am. Soc. C. E., was then in charge of the party of which 
Mr. Schofield was a member. 

Late in the fall of 1869 he returned to Rhode Island and re-entered 
the office of the late Samuel B. Gushing, Sr., M. Am. Soc. G. E., in 
Providence, where, with the exception of about two years spent on the 
Northern Pacific Railroad, he remained until the death of Samuel B. 
Gushing, Jr., M. Am. Soc. G. E., which occurred in 1888. He then 
carried on the business as the successor of the Gushings, until his 
death in November, 1908. 

In 1869 and 1870 he was leveler on the preliminary survey for the 
Milford and Lowell Railroad, of which the elder Mr. Gushing was 
Ghief Engineer. In 1873-1874 he was Engineer in charge of the con- 
struction of the East Providence branch of the Providence and Wor- 
cester Railroad. In the spring of 1881 Mr. Schofield again went West, 
and, until September, 1882, was Assistant Engineer on the Yellowstone 
Division of the Northern Pacific Railroad, his section lying between 
Billings and Miles Gity, Montana. Here he served with great credit, 
and, by his faithful and painstaking work, won the confidence and 
respect of those in the Ghief's office. After the completion of the 
Northern Pacific Railroad, Mr. Schofield resided in Providence and 
conducted a conservative engineering business of a general nature, 
doing active work up to within three weeks of the time of his death. 
*Memoir prepared by W. H. G. Temple. M. Am. Soc. C. E. 



MEMOIR OF MARK WILLIAM SCHOFIELD 433 

His whole life was marked by that sterling character, unswerving 
honesty, and strict loyalty to the interests of his clients, which won the 
respect of all with whom he had either business or social relations, and 
he unquestionably filled the essential requirements of an honorable 
man. 

Mr. Schofield was married on December 18th, 1873, to Annie S. 
Brown, a descendant of Chad Brown, one of the early landowners of 
Providence. His widow, together with four children, survives him. 

Mr. Schofield was elected a Member of the American Society of 
Civil Engineeers on May 1st, 1907. 



TR ANSACTI ONS 



American Society of Civil Engineers 



INDEX 
VOLUME EXIII 

JUNE, 1909 



Subject Index, Page 430 
Author Index, Page 439 



Titles of papers are in quotation marks when given with the 
author's name. 



VOLUME LXIII 



SUBJECT INDEX 



BRIDGES. 

" Nickel Steel for —." J. A. L. Waddell. loi. Discussion: Charles 
Evan Fowler, M. F. Brown, H. P. Bell, L. J. Le Conte, W. K. Halt, 
John C. Ostrup, T. Claxton Fidler, Robert E. Johnston. Albert 
Lucius, G. Lindenthal, Henry S. Prichard, Henry Le Chatelier, A. 
Ross, L. Dumas, Victor Prittie Perry, W. H. Warren, William R. 
Webster, William H. Breithaupt, E. A. Stone, C. Codron, W. W. K. 
Sparrow, B. J. Lambert, William Marriott, Henry Rohwer, Samuel 
Tobias Wagner, A. W. Carpenter, Leon S. Moisseiff, James C. Hall- 
sted, F. Arnodin, Wilson Worsdell, and William F. Pettigrevv, 300. 

BUILDINGS. 

" Foundations for the New Singer Building, New York City.'' T. Ken- 
nard Thomson. (With Discussion.) i. 

CAISSONS. 

"Foundations for the New Singer Building, New York City." T. Ken- 
nard Thomson. (With Discussion.) i. 

COLUMNS. 

"Nickel Steel for Bridges." J. A. L. Waddell. (With Discussion.) loi. 
COMPENSATING WORKS. 

" The Low Stage of Lakes Huron and Michigan." C. E. Grunsky. 
(With Discussion.) 31. 

DAMS. 

"The Improvement of the Ohio River." William L. Sibert. (With Dis- 
cussion.) 388. 

ELECTRIC RAILWAYS. 

" — in the Ohio Valley between Steubenville, Ohio, and Vanport, Penn- 
sylvania." George B. Francis. 73. Discussion: F. Lavis, George 
B. Preston, J. Martin Schreiber, and William J. Boucher, g\. 

EYE-BARS. 

"Nickel Steel for Bridges." J. A. L. Waddell. (With Discussion.) loi. 

FLOODS. 

" The — of the Mississippi Delta: Their Causes, and Suggestions as to 
Their Contiol." William D. Pickett. 53. 



SUBJECT INDEX 437 

FORESTS. 

" The Floods of the Mississippi Delta : Their Causes, and Suggestions as 
to Their Control." William D. Pickett. 53. 

FOUNDATIONS. 

" — for the New Singer Building, New York City." T. Kennard Thom- 
son. I. Discussion : O. F. Semsch, Eugene W. Stern, and Edwin 
S. Jarrett, 21. 

LAKES. 

"The Low Stage of Lakes Huron and Michigan." C. E. Grunsky. 31. 
Discussion : H. M. Chittenden, 48. 

LOCKS. 

"The Improvement of the Ohio River." William L. Siliert. (With Dis- 
cussion.) 388. 

MEMOIRS OF DECEASED MEMBERS. 

Chase, William Beverly. 429. 
Mansfield, Martin William. 431. 
Schofield, Mark William. 432. 

MJETALLURGY. 

"Nickel Steel for Bridges. " J. A. L. Waddell. (With Discussion.) loi. 
MOVABLE DAMS. 

"The Improvement of the Ohio River." William L. Sibert. (With 
Discussion.) 388. 

NAVIGATION. 

" The Low Stage of Lakes Huron and Michigan." C. E. Crnnsky. 
(With Discussion.) 31. 

RAILROADS. 

" Electric Railways in the Ohio Valley between Steubenville, Ohio, and 
Vanport, Pennsylvania." George B. Francis. (With Discussion.) 11. 

RAILS. 

Form of — for electric railways. 96. 

RIVERS. 

" The Floods of the Mississippi Delta : Their Causes, and Suggestions 
as to Their Control." William D. Pickett. 53. 

'• The Improvement of the Ohio River." William L. Sibert. 388. Dis- 
cussion: Theron M. Ripley, 426. 

SNOW. 

" The Floods of the Mississippi Delta : Their Causes, and Suggestions 
as to Their Control." William D. Pickett. 53. 



438 SUBJECT INDEX 

STEEL. 

"Nickel — for Bridges." J. A. L. Waddell. (With Discussion.) loi. 
STRENGTH OF MATERIALS, 

" Nickel Steel for Bridges." J. A. L. Waddell. (With Discussion.) loi. 
WATER, FLOW OF, IN OPEN CHANNELS. 

"The Low Stage of Lakes Huron and Michigan." C. E. Grunsky. 
(With Discussion.) 31. 



\ 



I 



AUTHOR INDEX 

ARNODIN, F. 

Nickel steel for bridges. 366. 

BELL, H. P. 

Nickel steel for bridges. 302. 
BOUCHER, WILLIAM J. 

Electric railways in the Ohio Valley. 98. 
BREITHAUPT, WILLIAM H. 

Nickel steel for bridges. 339. 
BROWN, M. F. 

Nickel steel for bridges. 301. 
CARPENTER, A. W. 

Nickel steel for bridges. 352. 
CHASE, WILLIAM BEVERLY. 

Memoir of. 429. 

CHITTENDEN, H. M. 

Low stage of Lakes Huron and Michigan. 48. 
CODRON, C. 

Nickel steel for bridges. 342. 
DUMAS, L. 

Nickel steel for bridges. 327. 
FIDLER, T. CLAXTON. 

Nickel steel for bridges. 312. 
FOWLER, CHARLES EVAN. 

Nickel steel for bridges. 300. 
FRANCIS, GEORGE B. 

" Electric Railways in the Ohio Valley Iietween Steubenville, Ohio, and 
Vanport, Pennsylvania." 73. 

GRUNSKY, C. E. 

" The Low .Stage of Lakes Huron and Michigan." 31. 
HALLSTED, JAMES C. 

Nickel steel for bridges. 361. 



440 AUTHOR INDEX 

HATT, W. K. 

Nickel steel for bridges. 306. 
JARRETT, EDWIN S. 

Pneumatic foundations. 26. 
JOHNSTON, ROBERT E. 

Nickel steel for bridges. 315. 
LAMBERT, B. J. 

Nickel steel for bridges. 347. 
LAVIS, F. 

Electric railways in the Ohio Valley. 91. 
LE CHATELIER, HENRY. 

Nickel steel for bridges. 322. 
LE CONTE, L. J. 

Nickel steel for bridges. 305. 
LINDENTHAL, G. 

Nickel steel for bridges. 316, 
LUCIUS, ALBERT. 

Nickel steel for bridges. 315. 

MANSFIELD, MARTIN WILLIAM. 

Memoir of. 431. 

MARRIOTT, WILLIAM. 

Nickel steel for bridges. 348. 
MOISSEIFF, LEON S. 

Nickel steel for bridges. 358. 
OSTRUP, JOHN C. 

Nickel steel for bridges. 308. 

PERRY, VICTOR PRITTIE. 

Nicke! steel for bridges. 328. 
PETTIGREW, WILLIAM F. 

Nickel steel for bridges. 379. 
PICKETT, WILLIAM D. 

" The Floods of the Mississippi Delta: Their Causes, and Suggestions as 
to Their Control." 53. 

PRESTON, GEORGE B. 

Electric railways in the Ohio Valley. 96. 



AUTHOR INDEX 441 

PRICHARD, HENRY S. 

Nickel steel for bridges. 317. 
RIPLEY, THERON M. 

Improvement of the Ohio River. 426. 
ROHWER, HENRY. 

Nickel steel for bridges. 348. 
ROSS, A. 

Nickel steel for bridges. 324. 
SCHOFIELD, MARK WILLIAM. 

Memoir of. 432. 
SCHREIBER, J. MARTIN. 

P^lectric railways in the Ohio Valley. 96. 
SEMSCH, O. F. 

Pneumatic foundations. 21. 
SIBERT, WILLIAM L. 

" The Improvement of the Ohio River." 388. 
SPARROW, W, W. K. 

Nickel steel for bridges. 345. 
STERN, EUGENE W. 

Pneumatic foundations. 24. 
STONE, E. A. 

Nickel steel for bridges. 341. 

tho'mson, t. kennard. 

" Foundations for the New Singer Building, New York City." i. 
WADDELL, J. A. L. 

" Nickel Steel for Bridges." loi. 
WAGNER, SAMUEL TOBIAS. 

Nickel steel for bridges. 350. 
WARREN, W. H. 

Nickel steel for bridges. 331. 
WEBSTER. WILLIAM R, 

Nickel steel for bridges. 337. 
WORSDELL, WILSON. 

Nickel steel for bridges. 378. 



NORTHEASTERN UNIVERSITY LIBRARIES 

- ■ ilillllllllll!!! i ■-■"■""-'">"'i" 



3 9358 00860178 



NORTHEASTERN UNIVERSITY LIBRARIES 



3 9358 00860178